diff --git a/data/unified_datasets/README.md b/data/unified_datasets/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..a22a057e191e4bc2b75c0290e2d9cc8a09c23ffc
--- /dev/null
+++ b/data/unified_datasets/README.md
@@ -0,0 +1,561 @@
+# Unified data format with example
+
+Under `data/unified_datasets` directory.
+
+single turn->dialogue with one turn
+
+Each dataset have at least 4 files:
+
+- `README.md`: dataset description and the main changes from original data to processed data.
+- `preprocess.py`: python script that preprocess the data. By running `python preprocess.py` we can get the following two files. The structure `preprocess.py` should be:
+
+```python
+def preprocess():
+ pass
+if __name__ == '__main__':
+ preprocess()
+```
+
+- `ontology.json`: dataset ontology, contains descriptions, state definition, etc.
+- `data.json.zip`: contains `data.json`.
+
+### README
+
+- Data source: publication, original data download link, etc.
+- Data description:
+ - Annotations: whether have dialogue act, belief state annotation.
+ - Statistics: \# domains, # dialogues, \# utterances, Avg. turns, Avg. tokens (split by space), etc.
+- Main changes from original data to processed data.
+
+### Ontology
+
+`ontology.json`: a *dict* containing:
+
+- `domains`: (*dict*) descriptions for domains, slots. Must contains all slots in the state and non-binary dialogue acts.
+ - `$domain_name`: (*dict*)
+ - `description`: (*str*) description for this domain.
+ - `slots`: (*dict*)
+ - `$slot_name`: (*dict*)
+ - `description`: (*str*) description for this slot.
+ - `is_categorical`: (*bool*) categorical slot or not.
+ - `possible_values`: (*list*) List of possible values the slot can take. If the slot is a categorical slot, it is a complete list of all the possible values. If the slot is a non categorical slot, it is either an empty list or a small sample of all the values taken by the slot.
+
+- `intents`: (*dict*) descriptions for intents.
+ - `$intent_name`: (*dict*)
+ - `description`: (*str*) description for this intent.
+- `binary_dialogue_act`: (*list* of *dict*) special dialogue acts that the value may not present in the utterance, e.g. request the address of a hotel.
+ - `{"intent": (str), "domain": (str), "slot": (str), "value": (str)}`. domain, slot, value may be empty.
+- `state`: (*dict*) belief state of all domains.
+ - `$domain_name`: (*dict*)
+ - `$slot_name: ""`: slot with empty value. Note that the slot set are the subset of the slot set in Part 1 definition.
+
+### Dialogues
+
+`data.json`: a *list* of dialogues containing:
+
+- `dataset`: (*str*) dataset name, must be one of ['schema', 'multiwoz', 'camrest', 'woz', ...], and be the same as the current dataset.
+- `data_split`: (*str*) in [train, val, test].
+- `dialogue_id`: (*str*) use dataset name as prefix, add count.
+- `domains`: (*list*) domains in this dialogue.
+- `turns`: (*list* of *dict*)
+ - `speaker`: (*str*) "user" or "system". **User side first, user side final**, "user" and "system" appear alternately?
+ - `utterance`: (*str*) sentence.
+ - `utt_idx`: (*int*) `turns['utt_idx']` gives current turn.
+ - `dialogue_act`: (*dict*)
+ - `categorical`: (*list* of *dict*) for categorical slots.
+ - `{"intent": (str), "domain": (str), "slot": (str), "value": (str)}`. Value sets are defined in the ontology.
+ - `non-categorical` (*list* of *dict*) for non-categorical slots.
+ - `{"intent": (str), "domain": (str), "slot": (str), "value": (str), "start": (int), "end": (int)}`. `start` and `end` are character indexes for the value span.
+ - `binary` (*list* of *dict*) for binary dialogue acts in ontology.
+ - `{"intent": (str), "domain": (str), "slot": (str), "value": (str)}`. Possible dialogue acts are listed in the `ontology['binary_dialogue_act']`.
+ - `state`: (*dict*, optional, user side) full state are shown in `ontology['state']`.
+ - `$domain_name`: (*dict*) contains all slots in this domain.
+ - `$slot_name`: (*str*) value for this slot.
+ - `state_update`: (*dict*, optional, user side) records the difference of states between the current turn and the last turn.
+ - `categorical`: (*list* of *dict*) for categorical slots.
+ - `{"domain": (str), "slot": (str), "value": (str)}`. Value sets are defined in the ontology (**dontcare** may not be included).
+ - `non-categorical` (*list* of *dict*) for non-categorical slots.
+ - `{"domain": (str), "slot": (str), "value": (str), "utt_idx": (int), "start": (int), "end": (int)}`. `utt_idx` is the utterance index of the value. `start` and `end` are character indexes for the value span in the current turn. `turn[utt_idx]['utterance'][start:end]` gives the value.
+
+Other attributes are optional.
+
+Run `python evaluate.py $dataset` to check the validation of processed dataset.
+
+## Example of Schema Dataset
+
+```json
+ {
+ "dataset": "schema",
+ "data_split": "train",
+ "dialogue_id": "schema_535",
+ "original_id": "5_00022",
+ "domains": [
+ "event_2"
+ ],
+ "turns": [
+ {
+ "speaker": "user",
+ "utterance": "I feel like going out to do something in Oakland. I've heard the Raiders Vs Bengals game should be good.",
+ "utt_idx": 0,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "inform_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "geteventdates"
+ }
+ ],
+ "categorical": [],
+ "non-categorical": [
+ {
+ "intent": "inform",
+ "domain": "event_2",
+ "slot": "event_name",
+ "value": "raiders vs bengals",
+ "start": 65,
+ "end": 83
+ },
+ {
+ "intent": "inform",
+ "domain": "event_2",
+ "slot": "city",
+ "value": "oakland",
+ "start": 41,
+ "end": 48
+ }
+ ]
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "",
+ "time": "",
+ "number_of_tickets": "",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": [
+ {
+ "domain": "event_2",
+ "slot": "city",
+ "value": "oakland",
+ "utt_idx": 0,
+ "start": 41,
+ "end": 48
+ },
+ {
+ "domain": "event_2",
+ "slot": "event_name",
+ "value": "raiders vs bengals",
+ "utt_idx": 0,
+ "start": 65,
+ "end": 83
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "The Raiders Vs Bengals game is at Oakland-Alameda County Coliseum today.",
+ "utt_idx": 1,
+ "dialogue_act": {
+ "binary": [],
+ "categorical": [],
+ "non-categorical": [
+ {
+ "intent": "offer",
+ "domain": "event_2",
+ "slot": "date",
+ "value": "today",
+ "start": 66,
+ "end": 71
+ },
+ {
+ "intent": "offer",
+ "domain": "event_2",
+ "slot": "event_name",
+ "value": "raiders vs bengals",
+ "start": 4,
+ "end": 22
+ },
+ {
+ "intent": "offer",
+ "domain": "event_2",
+ "slot": "venue",
+ "value": "oakland-alameda county coliseum",
+ "start": 34,
+ "end": 65
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "What time does it start?",
+ "utt_idx": 2,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "time",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "",
+ "time": "",
+ "number_of_tickets": "",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "It starts at 7 pm.",
+ "utt_idx": 3,
+ "dialogue_act": {
+ "binary": [],
+ "categorical": [],
+ "non-categorical": [
+ {
+ "intent": "inform",
+ "domain": "event_2",
+ "slot": "time",
+ "value": "7 pm",
+ "start": 13,
+ "end": 17
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "That sounds fine.",
+ "utt_idx": 4,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "select",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "today",
+ "time": "",
+ "number_of_tickets": "",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": [
+ {
+ "domain": "event_2",
+ "slot": "date",
+ "value": "today",
+ "utt_idx": 1,
+ "start": 66,
+ "end": 71
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "Do you want to get tickets for it?",
+ "utt_idx": 5,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "offer_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "Yes, can you buy 3 tickets for me?",
+ "utt_idx": 6,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "affirm_intent",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [
+ {
+ "intent": "inform",
+ "domain": "event_2",
+ "slot": "number_of_tickets",
+ "value": "3"
+ }
+ ],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "today",
+ "time": "",
+ "number_of_tickets": "3",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [
+ {
+ "domain": "event_2",
+ "slot": "number_of_tickets",
+ "value": "3"
+ }
+ ],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "Sure. I will go ahead and buy 3 tickets for the Raiders Vs Bengals game in Oakland today. Is that right?",
+ "utt_idx": 7,
+ "dialogue_act": {
+ "binary": [],
+ "categorical": [
+ {
+ "intent": "confirm",
+ "domain": "event_2",
+ "slot": "number_of_tickets",
+ "value": "3"
+ }
+ ],
+ "non-categorical": [
+ {
+ "intent": "confirm",
+ "domain": "event_2",
+ "slot": "event_name",
+ "value": "raiders vs bengals",
+ "start": 48,
+ "end": 66
+ },
+ {
+ "intent": "confirm",
+ "domain": "event_2",
+ "slot": "date",
+ "value": "today",
+ "start": 83,
+ "end": 88
+ },
+ {
+ "intent": "confirm",
+ "domain": "event_2",
+ "slot": "city",
+ "value": "oakland",
+ "start": 75,
+ "end": 82
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "Yes, that's good. What's the address?",
+ "utt_idx": 8,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "venue_address",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "today",
+ "time": "",
+ "number_of_tickets": "3",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "The game is at 7000 Coliseum Way. I've bought the tickets.",
+ "utt_idx": 9,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "notify_success",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": [
+ {
+ "intent": "inform",
+ "domain": "event_2",
+ "slot": "venue_address",
+ "value": "7000 coliseum way",
+ "start": 15,
+ "end": 32
+ }
+ ]
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "Thanks! That's all.",
+ "utt_idx": 10,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "thank_you",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "today",
+ "time": "",
+ "number_of_tickets": "3",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "system",
+ "utterance": "Need help with anything else?",
+ "utt_idx": 11,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "req_more",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ }
+ },
+ {
+ "speaker": "user",
+ "utterance": "No, thank you.",
+ "utt_idx": 12,
+ "dialogue_act": {
+ "binary": [
+ {
+ "intent": "negate",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "thank_you",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "categorical": [],
+ "non-categorical": []
+ },
+ "state": {
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "raiders vs bengals",
+ "date": "today",
+ "time": "",
+ "number_of_tickets": "3",
+ "city": "oakland",
+ "venue": "",
+ "venue_address": ""
+ }
+ },
+ "state_update": {
+ "categorical": [],
+ "non-categorical": []
+ }
+ }
+ ]
+ }
+```
+
diff --git a/data/unified_datasets/camrest/README.md b/data/unified_datasets/camrest/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..ae06dbe9dd5cd62c980f6b9bc731a1542bfd1aad
--- /dev/null
+++ b/data/unified_datasets/camrest/README.md
@@ -0,0 +1,24 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 406 | 2936 | 7.23 | 11.36 | 1 |
+| dev | 135 | 941 | 6.97 | 11.99 | 1 |
+| train | 135 | 935 | 6.93 | 11.87 | 1 |
+
+## Main changes
+
+- domain is set to **restaurant**
+- ignore some rare pair
+- 3 values are not found in original utterances
+- **dontcare** values in non-categorical slots are calculated in `evaluate.py` so `da_match` in evaluation is lower than actual number.
+
+## Original data
+
+camrest used in convlab2, included in `data/` path
\ No newline at end of file
diff --git a/data/unified_datasets/camrest/data.zip b/data/unified_datasets/camrest/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..b0d3db9f816f377f431e33d4a43ab0b9eb668f2a
Binary files /dev/null and b/data/unified_datasets/camrest/data.zip differ
diff --git a/data/unified_datasets/camrest/ontology.json b/data/unified_datasets/camrest/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..b5617f7ef4d441674a361a280acdf3123468d53a
--- /dev/null
+++ b/data/unified_datasets/camrest/ontology.json
@@ -0,0 +1,122 @@
+{
+ "domains": {
+ "restaurant": {
+ "description": "find a restaurant to eat",
+ "slots": {
+ "food": {
+ "description": "food type the restaurant serves",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area where the restaurant is located",
+ "is_categorical": true,
+ "possible_values": [
+ "north",
+ "east",
+ "west",
+ "south",
+ "centre"
+ ]
+ },
+ "name": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price range of the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "cheap",
+ "moderate",
+ "expensive"
+ ]
+ },
+ "phone": {
+ "description": "phone number of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "exact location of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postal code of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "inform user of value of a slot"
+ },
+ "request": {
+ "description": "ask for value of a slot"
+ },
+ "nooffer": {
+ "description": "inform user that no restaurant matches his request"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "food",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "name",
+ "value": ""
+ }
+ ],
+ "state": {
+ "restaurant": {
+ "pricerange": "",
+ "area": "",
+ "food": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/camrest/original_data.zip b/data/unified_datasets/camrest/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..ab07af894f6508a78e7baf21d914978fe19e16a3
Binary files /dev/null and b/data/unified_datasets/camrest/original_data.zip differ
diff --git a/data/unified_datasets/camrest/preprocess.py b/data/unified_datasets/camrest/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..510df266aa4d528bfe19fd5c496be5d2f96521cb
--- /dev/null
+++ b/data/unified_datasets/camrest/preprocess.py
@@ -0,0 +1,350 @@
+import zipfile
+import json
+import os
+import copy
+import logging
+
+logging.basicConfig(level=logging.INFO)
+import sys
+sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
+# print(sys.path[-1])
+
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+
+self_dir = os.path.dirname(os.path.abspath(__file__))
+
+cat_slot_values = {
+ 'area': ['north', 'east', 'west', 'south', 'centre'],
+ 'pricerange': ['cheap', 'moderate', 'expensive']
+}
+
+camrest_desc = {
+ 'restaurant': {
+ 'domain': 'find a restaurant to eat',
+ 'food': 'food type the restaurant serves',
+ 'area': 'area where the restaurant is located',
+ 'name': 'name of the restaurant',
+ 'pricerange': 'price range of the restaurant',
+ 'phone': 'phone number of the restaurant',
+ 'address': 'exact location of the restaurant',
+ 'postcode': 'postal code of the restaurant',
+ },
+ 'intents': {
+ 'inform': 'inform user of value of a slot',
+ 'request': 'ask for value of a slot',
+ 'nooffer': 'inform user that no restaurant matches his request'
+ }
+}
+
+all_slots = ['food', 'area', 'name', 'pricerange', 'phone', 'address', 'postcode']
+
+
+def convert_da(utt, da, all_intent, all_binary_das):
+ converted_da = {
+ 'binary': [],
+ 'categorical': [],
+ 'non-categorical': []
+ }
+
+ for _intent, svs in da.items():
+ if _intent not in all_intent:
+ all_intent.append(_intent)
+
+ if _intent == 'nooffer':
+ converted_da['binary'].append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': '',
+ 'value': ''
+ })
+
+ if {
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': '',
+ 'value': ''
+ } not in all_binary_das:
+ all_binary_das.append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': '',
+ 'value': ''
+ })
+ continue
+
+ for s, v in svs:
+ if 'care' in v:
+ v = 'dontcare'
+ s = s.lower()
+ v = v.lower()
+ if _intent == 'request':
+ converted_da['binary'].append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': ''
+ })
+
+ if {
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': ''
+ } not in all_binary_das:
+ all_binary_das.append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': ''
+ })
+ continue
+
+ if s in cat_slot_values:
+ assert v in cat_slot_values[s] + ['dontcare']
+ converted_da['categorical'].append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v
+ })
+
+ else:
+ # non-categorical
+ start_ch = utt.find(v)
+
+ if start_ch == -1:
+ # if not v == 'dontcare':
+ # logging.info('non-categorical slot value not found')
+ # logging.info('value: {}'.format(v))
+ # logging.info('sentence: {}'.format(utt))
+ # continue
+
+ converted_da['non-categorical'].append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v,
+ # 'start': 0,
+ # 'end': 0
+ })
+ continue
+
+ converted_da['non-categorical'].append({
+ 'intent': _intent,
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': utt[start_ch: start_ch + len(v)],
+ 'start': start_ch,
+ 'end': start_ch + len(v)
+ })
+ assert utt[start_ch: start_ch + len(v)] == v
+
+ return converted_da
+
+
+def convert_state(state, state_slots):
+ ret_state = {'restaurant': {k: '' for k in state_slots}}
+ for da in state:
+ if da['act'] != 'inform':
+ continue
+
+ for s, v in da['slots']:
+ s = s.lower()
+ v = v.lower()
+
+ if not s in all_slots:
+ logging.info('state slot {} not in all_slots!'.format(s))
+ continue
+
+ ret_state['restaurant'][s] = v
+
+ if s not in state_slots:
+ print(s)
+ raise
+
+ return ret_state
+
+
+def get_state_update(prev_state, cur_state, prev_turns, cur_user_da, dialog_id):
+ # cur_user_da: List of non-categorical slot-values
+ diff_state = {}
+ state_update = {'categorical': [], 'non-categorical':[]}
+ for s, v in cur_state.items():
+ if s in prev_state and prev_state[s] == v:
+ continue
+ diff_state[s] = v
+
+ for s, v in diff_state.items():
+ if v == '':
+ continue
+ if s in cat_slot_values:
+ assert v in cat_slot_values[s] + ['dontcare']
+ state_update['categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v,
+ })
+ else:
+ # non-categorical slot
+ found = False
+ for _usr_da in cur_user_da:
+ if _usr_da['slot'] == s and _usr_da['value'] == v :
+ found = True
+ if v != 'dontcare' and 'start' in _usr_da:
+ state_update['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v,
+ 'utt_idx': len(prev_turns),
+ 'start': _usr_da['start'],
+ 'end': _usr_da['end']
+ })
+ else:
+ state_update['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v,
+ })
+ if found:
+ continue
+
+ prev_sys_da = [] if len(prev_turns) == 0 else prev_turns[-1]['dialogue_act']['non-categorical']
+ for _sys_da in prev_sys_da:
+ if _sys_da['slot'] == s and _sys_da['value'] == v and 'start' in _sys_da:
+ if _sys_da['slot'] == s and _sys_da['value'] == v:
+ state_update['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v,
+ 'utt_idx': len(prev_turns) - 1,
+ 'start': _sys_da['start'],
+ 'end': _sys_da['end']
+ })
+ found = True
+
+ if not found:
+ state_update['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': s,
+ 'value': v
+ })
+
+ return state_update
+
+
+def preprocess():
+ original_zipped_path = os.path.join(self_dir, 'original_data.zip')
+ if not os.path.exists(original_zipped_path):
+ raise FileNotFoundError(original_zipped_path)
+ if not os.path.exists(os.path.join(self_dir, 'data.zip')) or not os.path.exists(
+ os.path.join(self_dir, 'ontology.json')):
+ # print('unzip to', new_dir)
+ # print('This may take several minutes')
+ archive = zipfile.ZipFile(original_zipped_path, 'r')
+ archive.extractall(self_dir)
+
+ all_data = []
+ all_intent = []
+ all_binary_das = []
+ all_state_slots = ['pricerange', 'area', 'food']
+
+ data_splits = ['train', 'val', 'test']
+ extract_dir = os.path.join(self_dir, 'original_data')
+
+ if not os.path.exists('data.zip') or not os.path.exists('ontology.json'):
+
+ dialog_id = 1
+ for data_split in data_splits:
+ data = json.load(open(os.path.join(self_dir, extract_dir, '{}.json'.format(data_split))))
+
+ for i, d in enumerate(data):
+
+ dialogue = d['dial']
+ converted_dialogue = {
+ 'dataset': 'camrest',
+ 'data_split': data_split,
+ 'dialogue_id': 'camrest_' + str(dialog_id),
+ 'original_id': d['dialogue_id'],
+ 'domains': ['restaurant'],
+ 'turns': []
+ }
+
+ prev_state = {'restaurant': {}}
+ for turn in dialogue:
+ usr_text = turn['usr']['transcript'].lower()
+ usr_da = turn['usr']['dialog_act']
+
+ sys_text = turn['sys']['sent'].lower()
+ sys_da = turn['sys']['dialog_act']
+
+ cur_state = convert_state(turn['usr']['slu'], all_state_slots)
+ cur_user_da = convert_da(usr_text, usr_da, all_intent, all_binary_das)
+
+ usr_turn = {
+ 'utt_idx': len(converted_dialogue['turns']),
+ 'speaker': 'user',
+ 'utterance': usr_text,
+ 'dialogue_act': cur_user_da,
+ 'state': copy.deepcopy(cur_state),
+ 'state_update': get_state_update(prev_state['restaurant'], cur_state['restaurant'], converted_dialogue['turns'], cur_user_da['non-categorical'], converted_dialogue['dialogue_id'])
+ }
+
+ sys_turn = {
+ 'utt_idx': len(converted_dialogue['turns'])+1,
+ 'speaker': 'system',
+ 'utterance': sys_text,
+ 'dialogue_act': convert_da(sys_text, sys_da, all_intent, all_binary_das),
+ }
+
+ prev_state = copy.deepcopy(cur_state)
+
+ converted_dialogue['turns'].append(usr_turn)
+ converted_dialogue['turns'].append(sys_turn)
+ if converted_dialogue['turns'][-1]['speaker'] == 'system':
+ converted_dialogue['turns'].pop(-1)
+ all_data.append(converted_dialogue)
+ dialog_id += 1
+
+ json.dump(all_data, open('./data.json', 'w'), indent=4)
+ write_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ os.remove('data.json')
+
+ new_ont = {
+ 'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': [],
+ 'state': {}
+ }
+
+ new_ont['state']['restaurant'] = {}
+ for ss in all_state_slots:
+ new_ont['state']['restaurant'][ss] = ''
+
+ for b in all_binary_das:
+ new_ont['binary_dialogue_act'].append(b)
+
+ for i in all_intent:
+ new_ont['intents'][i] = {'description': camrest_desc['intents'][i]}
+
+ new_ont['domains']['restaurant'] = {
+ 'description': camrest_desc['restaurant']['domain'],
+ 'slots': {}
+ }
+ for s in all_slots:
+ new_ont['domains']['restaurant']['slots'][s] = {
+ "description": camrest_desc['restaurant'][s],
+ "is_categorical": True if s in cat_slot_values else False,
+ "possible_values": cat_slot_values[s] if s in cat_slot_values else []
+ }
+ json.dump(new_ont, open(os.path.join(self_dir, './ontology.json'), 'w'), indent=4)
+
+
+ else:
+ all_data = read_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ new_ont = json.load(open(os.path.join(self_dir, './ontology.json'), 'r'))
+
+ return all_data, new_ont
+
+
+if __name__ == '__main__':
+ preprocess()
diff --git a/data/unified_datasets/evaluate.py b/data/unified_datasets/evaluate.py
new file mode 100644
index 0000000000000000000000000000000000000000..1c68f384db76c34a6e05a4e3993af227e2e3be3c
--- /dev/null
+++ b/data/unified_datasets/evaluate.py
@@ -0,0 +1,334 @@
+import json
+import os
+from copy import deepcopy
+
+special_values = ['dontcare', '']
+
+
+def check_ontology(name):
+ """
+ ontology: {
+ "domains": {
+ domain name: {
+ "description": domain description,
+ "slots": {
+ slot name: {
+ "description": slot description
+ // possible_values is empty iff is_categorical is False
+ "is_categorical": is_categorical,
+ "possible_values": [possible_values...]
+ }
+ }
+ }
+ },
+ "intents": {
+ intent name: {
+ "description": intent description
+ }
+ },
+ "binary_dialogue_act": {
+ [
+ {
+ "intent": intent name,
+ "domain": domain name
+ "slot": slot name,
+ "value": some value
+ }
+ ]
+ }
+ "state": {
+ domain name: {
+ slot name: ""
+ }
+ }
+ }
+ """
+ global special_values
+
+ ontology_file = os.path.join(f'{name}', 'ontology.json')
+ assert os.path.exists(ontology_file), f'ontology file should named {ontology_file}'
+ ontology = json.load(open(ontology_file))
+
+ # record issues in ontology
+ descriptions = {
+ # if each domain has a description
+ "domains": True,
+ "slots": True,
+ "intents": True,
+ }
+ for domain_name, domain in ontology['domains'].items():
+ if not domain['description']:
+ descriptions["domains"] = False
+ # if not domain_name in ontology['state']:
+ # print(f"domain '{domain_name}' not found in state")
+ for slot_name, slot in domain["slots"].items():
+ if not slot["description"]:
+ descriptions["slots"] = False
+ if slot["is_categorical"]:
+ assert slot["possible_values"]
+ slot['possible_values'] = list(map(str.lower, slot['possible_values']))
+ for value in special_values:
+ assert value not in slot['possible_values'], f'ONTOLOGY\tspecial value `{value}` should not present in possible values'
+
+ for intent_name, intent in ontology["intents"].items():
+ if not intent["description"]:
+ descriptions["intents"] = False
+
+ binary_dialogue_acts = set()
+ for bda in ontology['binary_dialogue_act']:
+ assert bda['intent'] is None or bda["intent"] in ontology['intents'], f'ONTOLOGY\tintent undefined intent in binary dialog act: {bda}'
+ binary_dialogue_acts.add(tuple(bda.values()))
+ ontology['bda_set'] = binary_dialogue_acts
+
+ assert 'state' in ontology, 'ONTOLOGY\tno state'
+ redundant_value = False
+ for domain_name, domain in ontology['state'].items():
+ assert domain_name in ontology['domains']
+ for slot_name, value in domain.items():
+ assert slot_name in ontology['domains'][domain_name]['slots']
+ if value:
+ redundant_value = True
+
+ if redundant_value:
+ print('ONTOLOGY: redundant value description in state')
+
+ # print('description existence:', descriptions, '\n')
+ for description, value in descriptions.items():
+ if not value:
+ print(f'description of {description} is incomplete')
+ return ontology
+
+
+def check_data(name, ontology):
+ global special_values
+
+ from zipfile import ZipFile
+ data_file = os.path.join(f'{name}', 'data.zip')
+ if not os.path.exists(data_file):
+ print('cannot find data.zip')
+ return
+
+ print('loading data')
+ with ZipFile(data_file) as zipfile:
+ with zipfile.open('data.json', 'r') as f:
+ data = json.load(f)
+
+ all_id = set()
+ splits = ['train', 'val', 'test']
+ da_values = 0
+ da_matches = 0
+ state_values = 0
+ state_matches = 0
+ distances = []
+ stat_keys = ['dialogues', 'utterances', 'tokens', 'domains']
+ stat = {
+ split: {
+ key: 0 for key in stat_keys
+ } for split in splits
+ }
+
+ # present for both non-categorical or categorical
+
+ for dialogue in data:
+ dialogue_id = dialogue['dialogue_id']
+ assert isinstance(dialogue_id, str), '`dialogue_id` is expected to be str type'
+ dialogue_id = str(dialogue_id)
+
+ assert dialogue['dataset'] == name, f'{dialogue_id}\tinconsistent dataset name: {dialogue["dataset"]}'
+
+ split = dialogue['data_split']
+ assert split in splits, f'unknown split: `{split}`'
+ cur_stat = stat[split]
+ cur_stat['dialogues'] += 1
+ try:
+ prefix, num = dialogue_id.split('_')
+ assert prefix == name
+ int(num) # try converting to int
+ except:
+ print(f'{dialogue_id}\twrong dialogue id format: {dialogue_id}')
+ raise Exception
+ assert dialogue_id not in all_id, f'multiple dialogue id: {dialogue_id}'
+ all_id.add(dialogue_id)
+
+ cur_domains = dialogue['domains']
+ assert isinstance(cur_domains, list), f'{dialogue_id}\t`domains` is expected to be list type, '
+ assert len(set(cur_domains)) == len(cur_domains), f'{dialogue_id}\trepeated domains'
+ cur_stat['domains'] += len(cur_domains)
+ cur_domains = set(cur_domains)
+ for domain_name in cur_domains:
+ assert domain_name in ontology['domains'], f'{dialogue_id}\tundefined current domain: {domain_name}'
+
+ turns = dialogue['turns']
+ cur_stat['utterances'] += len(turns)
+ assert turns, f'{dialogue_id}\tempty turn'
+
+ assert turns[0]['speaker'] == 'user', f'{dialogue_id}\tnot start with user role'
+ if ontology['state']:
+ # update cur_state with state_update every turn, and compare it with state annotation
+ cur_state = {
+ domain_name: deepcopy(ontology['state'][domain_name]) for domain_name in cur_domains
+ }
+ # check dialog act
+ for turn_id, turn in enumerate(turns):
+ assert turn['speaker'] in ['user', 'system'], f'{dialogue_id}:{turn_id}\tunknown speaker value: {turn["speaker"]}'
+ assert turn_id == turn['utt_idx'], f'{dialogue_id}:{turn_id}\twrong utt_idx'
+ if turn_id > 0:
+ assert turns[turn_id - 1]['speaker'] != turn['speaker'], f'{dialogue_id}:{turn_id}\tuser and system should speak alternatively'
+
+ utterance = turn['utterance']
+ cur_stat['tokens'] += len(utterance.strip().split(' '))
+ dialogue_acts = turn['dialogue_act']
+
+ # check domain-slot-value
+ # prefix: error prefix
+ def check_dsv(domain_name, slot_name, value, categorical, prefix):
+ assert domain_name in cur_domains or domain_name == 'booking', f'{prefix}\t{domain_name} not presented in current domains'
+ domain = ontology['domains'][domain_name]
+ assert slot_name in domain['slots'], f'{prefix}\t{slot_name} not presented in domain {domain_name}'
+ slot = domain['slots'][slot_name]
+ if categorical:
+ assert slot['is_categorical'], f'{prefix}\t{domain_name}-{slot_name} is not categorical'
+ value = value.lower()
+ assert value in special_values or value in slot['possible_values'], f'{prefix}\t`{value}` not presented in possible values of' \
+ f' {domain_name}-{slot_name}: {slot["possible_values"]}'
+ else:
+ assert not slot['is_categorical'], f'{prefix}\t{domain_name}-{slot_name} is not non-categorical'
+
+ def check_da(da, categorical):
+ assert da['intent'] in ontology['intents'], f'{dialogue_id}:{turn_id}\tundefined intent {da["intent"]}'
+ check_dsv(da['domain'], da['slot'], da['value'], categorical, f'{dialogue_id}:{turn_id}:da')
+
+ for da in dialogue_acts['categorical']:
+ check_da(da, True)
+ for da in dialogue_acts['non-categorical']:
+ check_da(da, False)
+ # values only match after .strip() in some case, it's the issue of pre-processing
+ if da['value'] not in special_values:
+ da_values += 1
+ assert 'start' in da and 'end' in da or 'start' not in da and 'end' not in da, \
+ f'{dialogue_id}:{turn_id}\tstart and end field in da should both present or neither not present'
+ if 'start' in da:
+ value = utterance[da['start']:da['end']]
+ if da['value'].lower() == value.lower():
+ da_matches += 1
+
+ for da in dialogue_acts['binary']:
+ assert tuple(da.values()) in ontology['bda_set'], f'{dialogue_id}:{turn_id}\tbinary dialog act {da} not present in ontology'
+ # do not check domain-slot-value in binary dialogue acts
+
+ if turn['speaker'] == 'user':
+ assert 'state' in turn and 'state_update' in turn, f"{dialogue_id}:{turn_id}\tstate and state_update must present in user's role"
+ state_update = turn['state_update']
+
+ def apply_update(update, categorical):
+ domain_name = update['domain']
+ slot_name = update['slot']
+ value = update['value']
+ check_dsv(domain_name, slot_name, value, categorical, f'{dialogue_id}:{turn_id}:state_update')
+ cur_state[domain_name][slot_name] = value
+ if ontology['state']:
+ for update in state_update['categorical']:
+ apply_update(update, True)
+ for update in state_update['non-categorical']:
+ apply_update(update, False)
+ value = update['value']
+ if value not in special_values:
+ state_values += 1
+ if 'utt_idx' in update:
+ if turns[update['utt_idx']]['utterance'][update['start']:update['end']].lower() == update['value']:
+ state_matches += 1
+ else:
+ print('value in utt:\t', turns[update['utt_idx']]['utterance'][update['start']:update['end']].strip())
+ print('value in state:\t', update['value'])
+ pass
+
+ assert cur_state == turn['state'], f'{dialogue_id}:{turn_id}:state_update incorrect state or state update calculation'
+
+ else:
+ assert 'state' not in turn or 'state_update' in turn, f"{dialogue_id}:{turn_id}\tstate or state_update cannot present in system's role"
+
+ assert turns[-1]['speaker'] == 'user', f'{dialogue_id} dialog must end with user role'
+
+ if da_values:
+ print('da values match rate: {:.3f}'.format(da_matches * 100 / da_values))
+ if state_values:
+ print('state values match rate: {:.3f}'.format(state_matches * 100 / state_values))
+
+ all_stat = {key: 0 for key in stat_keys}
+ for key in stat_keys:
+ all_stat[key] = sum(stat[split][key] for split in splits)
+ stat['all'] = all_stat
+
+ for split in splits + ['all']:
+ cur_stat = stat[split]
+ if cur_stat['dialogues']:
+ cur_stat['avg_utt'] = round(cur_stat['utterances'] / cur_stat['dialogues'], 2)
+ cur_stat['avg_tokens'] = round(cur_stat['tokens'] / cur_stat['utterances'], 2)
+ cur_stat['avg_domains'] = round(cur_stat.pop('domains') / cur_stat['dialogues'], 2)
+ else:
+ del stat[split]
+ print(f'domains: {len(ontology["domains"])}')
+ print(json.dumps(stat, indent=4))
+ if state_matches:
+ for dis, cnt in enumerate(distances):
+ print(cnt)
+
+
+if __name__ == '__main__':
+ from argparse import ArgumentParser
+
+ parser = ArgumentParser(description="evaluate pre-processed datasets")
+ parser.add_argument('datasets', metavar='dataset_name', nargs='*', help='dataset names to be evaluated')
+ parser.add_argument('--all', action='store_true', help='evaluate all datasets')
+ parser.add_argument('--no-int', action='store_true', help='not interrupted by exception')
+ parser.add_argument('--preprocess', '-p', action='store_true', help='run preprocess automatically')
+ args = parser.parse_args()
+
+ if args.all:
+ datasets = list(filter(os.path.isdir, os.listdir()))
+ else:
+ datasets = args.datasets
+ if not datasets:
+ print('no dataset specified')
+ parser.print_help()
+ exit(1)
+
+ print('datasets to be evaluated:', datasets)
+
+ fail = []
+
+ for name in datasets:
+ try:
+ print('')
+ if not os.path.isdir(name):
+ print(f'dataset {name} not found')
+ continue
+
+ print(f'checking {name}')
+ preprocess_file = os.path.join(f'{name}', 'preprocess.py')
+ if not os.path.exists(preprocess_file):
+ print('no preprocess.py')
+ if args.preprocess:
+ print(f'skip evaluation of {name}')
+ continue
+ if args.preprocess:
+ print('pre-processing')
+
+ os.chdir(name)
+ import importlib
+ preprocess = importlib.import_module(f'{name}.preprocess')
+ preprocess.preprocess()
+ os.chdir('..')
+
+ ontology = check_ontology(name)
+ check_data(name, ontology)
+ except Exception as e:
+ if args.no_int:
+ fail.append(name)
+ else:
+ raise e
+
+ if not fail:
+ print('all datasets passed test')
+ else:
+ print('failed dataset(s):', fail)
diff --git a/data/unified_datasets/frames/README.md b/data/unified_datasets/frames/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..bd1fc1ba33b1f19b8dc41ca6b361eb2478564cb7
--- /dev/null
+++ b/data/unified_datasets/frames/README.md
@@ -0,0 +1,23 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 1369 | 19445 | 14.2 | 12.71 | 1 |
+
+## Main changes
+
+- domain is set to **travel**
+- slot-value pair changes: intent-book => book-"True", action-book => booked-"True"
+- ignore some rare pair
+- not annotate state and state upadte
+- span info is from string matching, covering 96.4 non-categorical value
+
+## Original data
+
+https://www.microsoft.com/en-us/research/project/frames-dataset/#!download
\ No newline at end of file
diff --git a/data/unified_datasets/frames/data.zip b/data/unified_datasets/frames/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..d6b73c808d8682571e2cd060a2937c34abd06e4e
Binary files /dev/null and b/data/unified_datasets/frames/data.zip differ
diff --git a/data/unified_datasets/frames/ontology.json b/data/unified_datasets/frames/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..4101d15b77135162e4cad5af5d433a434ad5c61d
--- /dev/null
+++ b/data/unified_datasets/frames/ontology.json
@@ -0,0 +1,1488 @@
+{
+ "domains": {
+ "travel": {
+ "description": "Book a vacation package containing round-trip flights and a hotel.",
+ "slots": {
+ "dst_city": {
+ "description": "Destination city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "or_city": {
+ "description": "Origin city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "str_date": {
+ "description": "Start date for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "n_adults": {
+ "description": "Number of adults",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "budget": {
+ "description": "The amount of money that the user has available to spend for the trip.",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "flex": {
+ "description": "Boolean value indicating whether the constraints are flexible",
+ "is_categorical": true,
+ "possible_values": [
+ "false",
+ "true"
+ ]
+ },
+ "duration": {
+ "description": "Duration of the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "ref_anaphora": {
+ "description": "Words used to refer to a frame",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Price of the trip including flights and hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "end_date": {
+ "description": "End date for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "max_duration": {
+ "description": "Maximum number of days for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "Name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "category": {
+ "description": "Rating of the hotel (in number of stars)",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "wifi": {
+ "description": "Boolean value indicating whether or not the hotel offers free wifi",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "dep_time_or": {
+ "description": "Time of departure from origin city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "n_children": {
+ "description": "Number of children",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "gst_rating": {
+ "description": "Rating of the hotel by guests (in number of stars)",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "parking": {
+ "description": "Boolean value indicating whether or not the hotel offers free parking",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "arr_time_or": {
+ "description": "Time of arrival to origin city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "breakfast": {
+ "description": "Boolean value indicating whether or not the hotel offers free breakfast",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "Number of different packages",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "seat": {
+ "description": "Seat type (economy or business)",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count_name": {
+ "description": "Number of different hotels",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "budget_ok": {
+ "description": "Boolean value indicating whether the package fits the budget",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "arr_time_dst": {
+ "description": "Time of arrival to destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "dep_time_dst": {
+ "description": "Time of departure from destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "gym": {
+ "description": "Boolean value indicating whether or not the hotel offers gym",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "spa": {
+ "description": "Boolean value indicating whether or not the hotel offers spa",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "downtown": {
+ "description": "Boolean value indicating whether or not the hotel is in the heart of the city",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count_dst_city": {
+ "description": "Number of destination cities",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "min_duration": {
+ "description": "Minimum number of days for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "airport": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of an airport",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "beach": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a beach",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "museum": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a museum",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "theatre": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a theatre",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "park": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a park",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "market": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a market",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "shopping": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a shopping center",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "university": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of an university",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "mall": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a mall",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "palace": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a palace",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ },
+ "cathedral": {
+ "description": "Boolean value indicating whether or not the hotel is in the vicinity of a cathedral",
+ "is_categorical": true,
+ "possible_values": [
+ "true"
+ ]
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "Inform a slot value"
+ },
+ "suggest": {
+ "description": "Suggest a slot value or package that does not match the user's constraints"
+ },
+ "no_result": {
+ "description": "Tell the user that the database returned no results"
+ },
+ "negate": {
+ "description": "Negate something said by the other speaker"
+ },
+ "switch_frame": {
+ "description": "Switch to a frame"
+ },
+ "request": {
+ "description": "Ask for the value of a particular slot"
+ },
+ "affirm": {
+ "description": "Affirm something said by the other speaker"
+ },
+ "offer": {
+ "description": "Offer a package to the user"
+ },
+ "request_alts": {
+ "description": "Ask for other possibilities"
+ },
+ "request_compare": {
+ "description": "Ask the wizard to compare packages"
+ },
+ "confirm": {
+ "description": "Ask the other speaker to confirm a given slot value"
+ },
+ "hearmore": {
+ "description": "Ask the user if she'd like to hear more about a given package"
+ },
+ "moreinfo": {
+ "description": "Ask for more information on a given set of results"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "book",
+ "value": "true"
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "booked",
+ "value": "true"
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "seat",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "count",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "count_name",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": "true"
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "breakfast",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "gst_rating",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "affirm",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "book",
+ "value": "true"
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "booked",
+ "value": "true"
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "confirm",
+ "domain": "travel",
+ "slot": "budget",
+ "value": "dontcare"
+ },
+ {
+ "intent": "hearmore",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "hearmore",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "book",
+ "value": "true"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "booked",
+ "value": "true"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "budget",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "dep_time_or",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "arr_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "dep_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "arr_time_or",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "category",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "spa",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": "false"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": "false"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "parking",
+ "value": "dontcare"
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "airport",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "booked",
+ "value": "true"
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "book",
+ "value": "true"
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "shopping",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "arr_time_or",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "gst_rating",
+ "value": ""
+ },
+ {
+ "intent": "negate",
+ "domain": "travel",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "dep_time_or",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "dep_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "museum",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "no_result",
+ "domain": "travel",
+ "slot": "downtown",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "offer",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": "false"
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "breakfast",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "max_duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "seat",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "count_dst_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "count_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "dep_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "dep_time_or",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "arr_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "arr_time_or",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "wifi",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "gym",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "flex",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "parking",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "spa",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "gst_rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "count",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "beach",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "park",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "museum",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "downtown",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "mall",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "shopping",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "budget",
+ "value": "dontcare"
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "university",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel",
+ "slot": "palace",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "dep_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "wifi",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "parking",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "breakfast",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "seat",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "gst_rating",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "dep_time_or",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "arr_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "dep_time_dst",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "arr_time_or",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "request_compare",
+ "domain": "travel",
+ "slot": "beach",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "dst_city",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "booked",
+ "value": "true"
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "budget",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "budget",
+ "value": "dontcare"
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "amenities",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "vicinity",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "museum",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "park",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "beach",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "n_adults",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "n_children",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "market",
+ "value": ""
+ },
+ {
+ "intent": "suggest",
+ "domain": "travel",
+ "slot": "or_city",
+ "value": "dontcare"
+ },
+ {
+ "intent": "switch_frame",
+ "domain": "travel",
+ "slot": "museum",
+ "value": ""
+ },
+ {
+ "intent": "switch_frame",
+ "domain": "travel",
+ "slot": "str_date",
+ "value": ""
+ },
+ {
+ "intent": "switch_frame",
+ "domain": "travel",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "switch_frame",
+ "domain": "travel",
+ "slot": "gst_rating",
+ "value": ""
+ }
+ ],
+ "state": {}
+}
\ No newline at end of file
diff --git a/data/unified_datasets/frames/original_data.zip b/data/unified_datasets/frames/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..07ca92735ee93112b830e6181aa9832a091698d7
Binary files /dev/null and b/data/unified_datasets/frames/original_data.zip differ
diff --git a/data/unified_datasets/frames/preprocess.py b/data/unified_datasets/frames/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..64a61f0dcae4d14e5dcc4e8e377bc5d7699a7eeb
--- /dev/null
+++ b/data/unified_datasets/frames/preprocess.py
@@ -0,0 +1,341 @@
+import zipfile
+import json
+import os
+from pprint import pprint
+from copy import deepcopy
+from collections import Counter
+from tqdm import tqdm
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+import re
+self_dir = os.path.dirname(os.path.abspath(__file__))
+
+
+intent2des = {
+ "inform": "Inform a slot value",
+ "offer": "Offer a package to the user",
+ "request": "Ask for the value of a particular slot",
+ "switch_frame": "Switch to a frame",
+ "suggest": "Suggest a slot value or package that does not match the user's constraints",
+ "no_result": "Tell the user that the database returned no results",
+ "thankyou": "Thank the other speaker",
+ "sorry": "Apologize to the user",
+ "greeting": "Greet the other speaker",
+ "affirm": "Affirm something said by the other speaker",
+ "negate": "Negate something said by the other speaker",
+ "confirm": "Ask the other speaker to confirm a given slot value",
+ "moreinfo": "Ask for more information on a given set of results",
+ "goodbye": "Say goodbye to the other speaker",
+ "request_alts": "Ask for other possibilities",
+ "request_compare": "Ask the wizard to compare packages",
+ "hearmore": "Ask the user if she'd like to hear more about a given package",
+ "you_are_welcome": "Tell the user she is welcome",
+ "canthelp": "Tell the user you cannot answer her request",
+ "reject": "Tell the user you did not understand what she meant"
+}
+
+slot2des = {
+ "book": "Find a trip to book",
+ "dst_city": "Destination city",
+ "or_city": "Origin city",
+ "str_date": "Start date for the trip",
+ "n_adults": "Number of adults",
+ "budget": "The amount of money that the user has available to spend for the trip.",
+ "end_date": "End date for the trip",
+ "flex": "Boolean value indicating whether the constraints are flexible",
+ "duration": "Duration of the trip",
+ "ref_anaphora": "Words used to refer to a frame",
+ "price": "Price of the trip including flights and hotel",
+ "max_duration": "Maximum number of days for the trip",
+ "amenities": "Number of amenities",
+ "vicinity": "Vicinity of the hotel",
+ "name": "Name of the hotel",
+ "category": "Rating of the hotel (in number of stars)",
+ "wifi": "Boolean value indicating whether or not the hotel offers free wifi",
+ "booked": "Booked a trip",
+ "dep_time_or": "Time of departure from origin city",
+ "n_children": "Number of children",
+ "gst_rating": "Rating of the hotel by guests (in number of stars)",
+ "parking": "Boolean value indicating whether or not the hotel offers free parking",
+ "arr_time_or": "Time of arrival to origin city",
+ "breakfast": "Boolean value indicating whether or not the hotel offers free breakfast",
+ "count": "Number of different packages",
+ "seat": "Seat type (economy or business)",
+ "count_name": "Number of different hotels",
+ "count_dst_city": "Number of destination cities",
+ "budget_ok": "Boolean value indicating whether the package fits the budget",
+ "arr_time_dst": "Time of arrival to destination",
+ "dep_time_dst": "Time of departure from destination",
+ "gym": "Boolean value indicating whether or not the hotel offers gym",
+ "spa": "Boolean value indicating whether or not the hotel offers spa",
+ "downtown": "Boolean value indicating whether or not the hotel is in the heart of the city",
+ "min_duration": "Minimum number of days for the trip",
+ "airport": "Boolean value indicating whether or not the hotel is in the vicinity of an airport",
+ "beach": "Boolean value indicating whether or not the hotel is in the vicinity of a beach",
+ "museum": "Boolean value indicating whether or not the hotel is in the vicinity of a museum",
+ "theatre": "Boolean value indicating whether or not the hotel is in the vicinity of a theatre",
+ "park": "Boolean value indicating whether or not the hotel is in the vicinity of a park",
+ "market": "Boolean value indicating whether or not the hotel is in the vicinity of a market",
+ "shopping": "Boolean value indicating whether or not the hotel is in the vicinity of a shopping center",
+ "university": "Boolean value indicating whether or not the hotel is in the vicinity of an university",
+ "mall": "Boolean value indicating whether or not the hotel is in the vicinity of a mall",
+ "palace": "Boolean value indicating whether or not the hotel is in the vicinity of a palace",
+ "cathedral": "Boolean value indicating whether or not the hotel is in the vicinity of a cathedral",
+ "no_result": "Boolean value indicating whether there is no result match user's constraints"
+}
+
+
+def get_slot_type(slot):
+ if slot in {'book', 'booked', 'vicinity', 'amenities'}:
+ return 'binary'
+ elif slot in {'dst_city', 'or_city', 'str_date', 'end_date', 'duration', 'min_duration', 'max_duration',
+ 'dep_time_or', 'arr_time_or', 'arr_time_dst', 'dep_time_dst', 'n_adults', 'n_children', 'budget',
+ 'price', 'ref_anaphora', 'name', 'category', 'gst_rating',
+ 'count', 'count_name', 'count_dst_city', 'seat'}:
+ return 'non-categorical'
+ elif slot in {'budget_ok', 'flex', 'wifi', 'parking', 'breakfast', 'gym', 'spa', 'downtown', 'airport', 'beach',
+ 'museum', 'theatre', 'park', 'market', 'shopping', 'university', 'mall', 'palace', 'cathedral'}:
+ return 'categorical'
+ else:
+ return None
+
+
+digit2word = {
+ '0': 'zero', '1': 'one', '2': 'two', '3': 'three', '4': 'four', '5': 'five',
+ '6': 'six', '7': 'seven', '8': 'eight', '9': 'nine', '10': 'ten'
+}
+
+match = {
+ '0': 0,
+ '1': 0,
+ '>1': 0,
+}
+
+
+def pharse_in_sen(phrase, sen):
+ '''
+ match value in the sentence
+ :param phrase: str
+ :param sen: str
+ :return: start, end if matched, else None, None
+ '''
+ assert isinstance(phrase, str)
+ pw = '(^|[\s,\.:\?!-])(?P<v>{})([\s,\.:\?!-]|$)'
+ pn = '(^|[\s\?!-]|\D[,\.:])(?P<v>{})($|[\s\?!-]|[,\.:]\D|[,\.:]$)'
+ if phrase.isdigit():
+ pattern = pn
+ else:
+ pattern = pw
+ p = re.compile(pattern.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ if num > 1:
+ match['>1'] += 1
+ else:
+ match['1'] += 1
+ return m.span('v'), num
+ if phrase.isdigit() and phrase in digit2word:
+ phrase = digit2word[phrase]
+ p = re.compile(pw.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ if num > 1:
+ match['>1'] += 1
+ else:
+ match['1'] += 1
+ return m.span('v'), num
+ match['0'] += 1
+ return (None, None), 0
+
+
+def iter_over_acts(acts):
+ for act in acts:
+ intent = act['name']
+ for arg in act['args']:
+ k = arg['key']
+ if k == 'id':
+ continue
+ elif k in ['ref', 'read', 'write']:
+ assert isinstance(arg['val'], list)
+ for frame in arg['val']:
+ for kv in frame['annotations']:
+ if kv['key'] in ('ref', 'read', 'write'):
+ print(kv, frame)
+ assert False
+ yield intent, kv['key'], kv.get('val')
+ else:
+ yield intent, k, arg.get('val', None)
+
+
+def normalize_da(intent, slot, value, utterance):
+ if slot == 'intent':
+ slot = 'book'
+ elif slot == 'action':
+ slot = 'booked'
+ elif slot not in slot2des:
+ # ignore some rare slot
+ return None, None
+
+ if slot in ['book', 'booked']:
+ slot_type = 'binary'
+ return slot_type, {
+ "intent": intent,
+ "domain": 'travel',
+ "slot": slot,
+ "value": 'True',
+ }
+ elif value is None or value == '':
+ slot_type = 'binary'
+ return slot_type, {
+ "intent": intent,
+ "domain": 'travel',
+ "slot": slot,
+ "value": '',
+ }
+ elif value == '-1':
+ slot_type = 'binary'
+ return slot_type, {
+ "intent": intent,
+ "domain": 'travel',
+ "slot": slot,
+ "value": 'dontcare',
+ }
+ elif isinstance(value, str):
+ slot_type = get_slot_type(slot)
+ assert slot_type == 'non-categorical'
+ (start, end), num = pharse_in_sen(value, utterance)
+ if not num:
+ if slot == 'gst_rating' and pharse_in_sen(' / '.join(value.split('/')), utterance)[1]:
+ value = ' / '.join(value.split('/'))
+ elif 'a. m' in value and pharse_in_sen(value.replace('a. m', 'a.m'), utterance)[1]:
+ value = value.replace('a. m', 'a.m')
+ elif 'p. m' in value and pharse_in_sen(value.replace('p. m', 'p.m'), utterance)[1]:
+ value = value.replace('p. m', 'p.m')
+ elif slot == 'price' and pharse_in_sen(value.replace('USD', ' USD'), utterance)[1]:
+ value = value.replace('USD', ' USD')
+ else:
+ # few wrong annotation
+ return None, None
+ (start, end), num = pharse_in_sen(value, utterance)
+ assert num, print(value, utterance)
+ if not num:
+ return None, None
+ # return None, None
+ return slot_type, {
+ "intent": intent,
+ "domain": 'travel',
+ "slot": slot,
+ "value": utterance[start:end],
+ "start": start,
+ "end": end
+ }
+ elif isinstance(value, bool):
+ slot_type = get_slot_type(slot)
+ value = str(value)
+ assert slot_type == 'categorical' or slot_type == 'binary', print(slot, value)
+ return slot_type, {
+ "intent": intent,
+ "domain": 'travel',
+ "slot": slot,
+ "value": value,
+ }
+ else:
+ assert 0
+
+
+def preprocess():
+ processed_dialogue = []
+ ontology = {'domains': {'travel':
+ {"description": "Book a vacation package containing round-trip flights and a hotel.",
+ "slots": {}}},
+ 'intents': {},
+ 'binary_dialogue_act': [],
+ 'state': {}}
+ original_zipped_path = os.path.join(self_dir, 'original_data.zip')
+ new_dir = os.path.join(self_dir, 'original_data')
+ if not os.path.exists(original_zipped_path):
+ raise FileNotFoundError(original_zipped_path)
+ if not os.path.exists(os.path.join(self_dir, 'data.zip')) or not os.path.exists(os.path.join(self_dir, 'ontology.json')):
+ print('unzip to', new_dir)
+ print('This may take several minutes')
+ archive = zipfile.ZipFile(original_zipped_path, 'r')
+ archive.extractall(new_dir)
+ data = json.load(open(os.path.join(new_dir, 'frames.json')))
+ # json.dump(data, open(os.path.join(new_dir, 'original_data.json'), 'w'), indent=2)
+ cnt = 1
+ for d in tqdm(data, desc='dialogue'):
+ dialogue = {
+ "dataset": 'frames',
+ "data_split": 'train',
+ "dialogue_id": 'frames_' + str(cnt),
+ "original_id": d['id'],
+ "user_id": d['user_id'],
+ "system_id": d['wizard_id'],
+ "userSurveyRating": d['labels']['userSurveyRating'],
+ "wizardSurveyTaskSuccessful": d['labels']['wizardSurveyTaskSuccessful'],
+ "domains": ['travel'],
+ "turns": []
+ }
+ # state = deepcopy(ontology['state']['travel'])
+ for utt_idx, t in enumerate(d['turns']):
+ speaker = 'system' if t['author']=='wizard' else t['author']
+ turn = {
+ 'speaker': speaker,
+ 'utterance': t['text'],
+ 'utt_idx': utt_idx,
+ 'dialogue_act': {
+ 'binary': [],
+ 'categorical': [],
+ 'non-categorical': [],
+ },
+ }
+ for intent, slot, value in iter_over_acts(t['labels']['acts']):
+ da_type, da = normalize_da(intent, slot, value, t['text'])
+ if da is not None:
+ da['value'] = da['value'].lower()
+ turn['dialogue_act'][da_type].append(da)
+ slot = da['slot']
+ value = da['value']
+ if da_type == 'binary':
+ if da not in ontology['binary_dialogue_act']:
+ ontology['binary_dialogue_act'].append(da)
+ else:
+ ontology['domains']['travel']['slots'].setdefault(slot, {
+ "description": slot2des[slot],
+ "is_categorical": da_type=='categorical',
+ "possible_values": []
+ })
+ if da_type == 'categorical' \
+ and value not in ontology['domains']['travel']['slots'][slot]['possible_values']:
+ ontology['domains']['travel']['slots'][slot]['possible_values'].append(value)
+ ontology['intents'].setdefault(intent, {
+ "description": intent2des[intent]
+ })
+ # state
+ if speaker == 'user':
+ turn['state'] = {}
+ turn['state_update'] = {
+ 'categorical': [],
+ 'non-categorical': [],
+ }
+ dialogue['turns'].append(deepcopy(turn))
+ cnt += 1
+ if len(dialogue['turns']) % 2 == 0:
+ dialogue['turns'] = dialogue['turns'][:-1]
+ processed_dialogue.append(deepcopy(dialogue))
+ ontology['binary_dialogue_act'] = sorted(ontology['binary_dialogue_act'], key=lambda x: x['intent'])
+ json.dump(ontology, open(os.path.join(self_dir, 'ontology.json'), 'w'), indent=2)
+ json.dump(processed_dialogue, open('data.json', 'w'), indent=2)
+ write_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ os.remove('data.json')
+ else:
+ # read from file
+ processed_dialogue = read_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ ontology = json.load(open(os.path.join(self_dir, 'ontology.json')))
+ return processed_dialogue, ontology
+
+
+if __name__ == '__main__':
+ preprocess()
+ print(match) # {'0': 271, '1': 29333, '>1': 806}
diff --git a/data/unified_datasets/metalwoz/README.md b/data/unified_datasets/metalwoz/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..e49e76f5a938161fa1f39f5f762173f5841aabe6
--- /dev/null
+++ b/data/unified_datasets/metalwoz/README.md
@@ -0,0 +1,17 @@
+# README
+
+## Features
+
+No sentence-level annotation. Only annotate domain.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 37884 | 362450 | 9.57 | 7.66 | - |
+| test | 2319 | 21949 | 9.46 | 8.23 | - |
+
+
+## Original data
+
+- https://www.microsoft.com/en-us/research/project/metalwoz/
diff --git a/data/unified_datasets/metalwoz/data.zip b/data/unified_datasets/metalwoz/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..ccd14f97ff4f8be6244d1634f33ca3b6ecda6dc8
Binary files /dev/null and b/data/unified_datasets/metalwoz/data.zip differ
diff --git a/data/unified_datasets/metalwoz/metalwoz-test-v1.zip b/data/unified_datasets/metalwoz/metalwoz-test-v1.zip
new file mode 100644
index 0000000000000000000000000000000000000000..040153c440767b113a805e68d7e9851e2e066c06
Binary files /dev/null and b/data/unified_datasets/metalwoz/metalwoz-test-v1.zip differ
diff --git a/data/unified_datasets/metalwoz/metalwoz-v1.zip b/data/unified_datasets/metalwoz/metalwoz-v1.zip
new file mode 100644
index 0000000000000000000000000000000000000000..eff7551b322eb7c3428d706fa660fce9c776aced
Binary files /dev/null and b/data/unified_datasets/metalwoz/metalwoz-v1.zip differ
diff --git a/data/unified_datasets/metalwoz/ontology.json b/data/unified_datasets/metalwoz/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..099a53528f47e5962b053c649d3e4a56cf0819c8
--- /dev/null
+++ b/data/unified_datasets/metalwoz/ontology.json
@@ -0,0 +1,211 @@
+{
+ "domains": {
+ "AGREEMENT_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "ALARM_SET": {
+ "description": "",
+ "slots": {}
+ },
+ "APARTMENT_FINDER": {
+ "description": "",
+ "slots": {}
+ },
+ "APPOINTMENT_REMINDER": {
+ "description": "",
+ "slots": {}
+ },
+ "AUTO_SORT": {
+ "description": "",
+ "slots": {}
+ },
+ "BANK_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "BUS_SCHEDULE_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "CATALOGUE_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "CHECK_STATUS": {
+ "description": "",
+ "slots": {}
+ },
+ "CITY_INFO": {
+ "description": "",
+ "slots": {}
+ },
+ "CONTACT_MANAGER": {
+ "description": "",
+ "slots": {}
+ },
+ "DECIDER_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "EDIT_PLAYLIST": {
+ "description": "",
+ "slots": {}
+ },
+ "EVENT_RESERVE": {
+ "description": "",
+ "slots": {}
+ },
+ "GAME_RULES": {
+ "description": "",
+ "slots": {}
+ },
+ "GEOGRAPHY": {
+ "description": "",
+ "slots": {}
+ },
+ "GUINESS_CHECK": {
+ "description": "",
+ "slots": {}
+ },
+ "HOME_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "HOW_TO_BASIC": {
+ "description": "",
+ "slots": {}
+ },
+ "INSURANCE": {
+ "description": "",
+ "slots": {}
+ },
+ "LIBRARY_REQUEST": {
+ "description": "",
+ "slots": {}
+ },
+ "LOOK_UP_INFO": {
+ "description": "",
+ "slots": {}
+ },
+ "MAKE_RESTAURANT_RESERVATIONS": {
+ "description": "",
+ "slots": {}
+ },
+ "MOVIE_LISTINGS": {
+ "description": "",
+ "slots": {}
+ },
+ "MUSIC_SUGGESTER": {
+ "description": "",
+ "slots": {}
+ },
+ "NAME_SUGGESTER": {
+ "description": "",
+ "slots": {}
+ },
+ "ORDER_PIZZA": {
+ "description": "",
+ "slots": {}
+ },
+ "PET_ADVICE": {
+ "description": "",
+ "slots": {}
+ },
+ "PHONE_PLAN_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "PHONE_SETTINGS": {
+ "description": "",
+ "slots": {}
+ },
+ "PLAY_TIMES": {
+ "description": "",
+ "slots": {}
+ },
+ "POLICY_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "PRESENT_IDEAS": {
+ "description": "",
+ "slots": {}
+ },
+ "PROMPT_GENERATOR": {
+ "description": "",
+ "slots": {}
+ },
+ "QUOTE_OF_THE_DAY_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "RESTAURANT_PICKER": {
+ "description": "",
+ "slots": {}
+ },
+ "SCAM_LOOKUP": {
+ "description": "",
+ "slots": {}
+ },
+ "SHOPPING": {
+ "description": "",
+ "slots": {}
+ },
+ "SKI_BOT": {
+ "description": "",
+ "slots": {}
+ },
+ "SPORTS_INFO": {
+ "description": "",
+ "slots": {}
+ },
+ "STORE_DETAILS": {
+ "description": "",
+ "slots": {}
+ },
+ "TIME_ZONE": {
+ "description": "",
+ "slots": {}
+ },
+ "UPDATE_CALENDAR": {
+ "description": "",
+ "slots": {}
+ },
+ "UPDATE_CONTACT": {
+ "description": "",
+ "slots": {}
+ },
+ "WEATHER_CHECK": {
+ "description": "",
+ "slots": {}
+ },
+ "WEDDING_PLANNER": {
+ "description": "",
+ "slots": {}
+ },
+ "WHAT_IS_IT": {
+ "description": "",
+ "slots": {}
+ },
+ "BOOKING_FLIGHT": {
+ "description": "",
+ "slots": {}
+ },
+ "HOTEL_RESERVE": {
+ "description": "",
+ "slots": {}
+ },
+ "TOURISM": {
+ "description": "",
+ "slots": {}
+ },
+ "VACATION_IDEAS": {
+ "description": "",
+ "slots": {}
+ }
+ },
+ "intents": {},
+ "binary_dialogue_act": [],
+ "state": {}
+}
\ No newline at end of file
diff --git a/data/unified_datasets/metalwoz/preprocess.py b/data/unified_datasets/metalwoz/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..c075e7d4541169a7fd9c611503d6ed9d8ae69817
--- /dev/null
+++ b/data/unified_datasets/metalwoz/preprocess.py
@@ -0,0 +1,89 @@
+import json
+import os
+from zipfile import ZipFile, ZIP_DEFLATED
+
+import json_lines
+
+
+dataset = 'metalwoz'
+self_dir = os.path.dirname(os.path.abspath(__file__))
+DATA_PATH = os.path.join(os.path.dirname(os.path.dirname(self_dir)), 'data')
+# origin_data_dir = os.path.join(DATA_PATH, dataset)
+origin_data_dir = self_dir
+
+
+def preprocess():
+ ontology = {
+ 'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': [],
+ 'state': {}
+ }
+
+ def process_dialog(ori_dialog, split, dialog_id):
+ domain = ori_dialog['domain']
+ ontology['domains'][domain] = {
+ 'description': "",
+ 'slots': {}
+ }
+ dialog = {
+ "dataset": dataset,
+ "data_split": split,
+ "dialogue_id": f'{dataset}_{dialog_id}',
+ "original_id": ori_dialog['id'],
+ "domains": [domain],
+ }
+ turns = []
+ # starts with system
+ for utt_idx, utt in enumerate(ori_dialog['turns'][1:]):
+ turn = {
+ 'utt_idx': utt_idx,
+ 'utterance': utt,
+ 'dialogue_act': {
+ 'categorical': [],
+ 'non-categorical': [],
+ 'binary': [],
+ },
+ }
+ if utt_idx % 2 == 0:
+ turn['speaker'] = 'user'
+ turn['state'] = {}
+ turn['state_update'] = {
+ 'categorical': [],
+ 'non-categorical': [],
+ }
+ else:
+ turn['speaker'] = 'system'
+ turns.append(turn)
+ if turns[-1]['speaker'] == 'system':
+ turns.pop()
+
+ dialog['turns'] = turns
+ return dialog
+
+ dialog_id = 0
+ data = []
+ with ZipFile(os.path.join(origin_data_dir, 'metalwoz-v1.zip')) as zipfile:
+ for path in zipfile.namelist():
+ if path.startswith('dialogues'):
+ for dialog in json_lines.reader(zipfile.open(path)):
+ data.append(process_dialog(dialog, 'train', dialog_id))
+ dialog_id += 1
+
+ ZipFile(os.path.join(origin_data_dir, 'metalwoz-test-v1.zip')).extract('dstc8_metalwoz_heldout.zip')
+ with ZipFile(os.path.join('dstc8_metalwoz_heldout.zip')) as zipfile:
+ for path in zipfile.namelist():
+ if path.startswith('dialogues'):
+ for dialog in json_lines.reader(zipfile.open(path)):
+ data.append(process_dialog(dialog, 'test', dialog_id))
+ dialog_id += 1
+ os.remove('dstc8_metalwoz_heldout.zip')
+
+ json.dump(ontology, open(os.path.join(self_dir, 'ontology.json'), 'w'))
+ json.dump(data, open('data.json', 'w'), indent=4)
+ ZipFile(os.path.join(self_dir, 'data.zip'), 'w', ZIP_DEFLATED).write('data.json')
+ os.remove('data.json')
+
+
+if __name__ == '__main__':
+ preprocess()
diff --git a/data/unified_datasets/multiwoz21/README.md b/data/unified_datasets/multiwoz21/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..803ebf3ad3ee6c4b6aaf710a739f9518bf5d5321
--- /dev/null
+++ b/data/unified_datasets/multiwoz21/README.md
@@ -0,0 +1,31 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots. state and state updates.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 8434 | 105066 | 12.46 | 17.27 | 7 |
+| dev | 999 | 13731 | 13.74 | 17.72 | 7 |
+| train | 1000 | 13744 | 13.74 | 17.67 | 7 |
+
+
+## Main changes
+
+- only keep 5 domains in state annotations and dialog acts.
+- `pricerange`, `area`, `day`, `internet`, `parking`, `stars` are considered categorical slots.
+- punctuation marks are split from their previous tokens. e.g `I want to find a hotel. ->
+ I want to find a hotel .`
+
+Run `evaluate.py`:
+
+da values match rate: 97.944
+state values match rate: 66.017
+
+### original data
+
+- from [multiwoz](https://github.com/budzianowski/multiwoz) repo.
+
diff --git a/data/unified_datasets/multiwoz21/data.zip b/data/unified_datasets/multiwoz21/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..7f60019dfc4a68728474baf9e8d3d73305b46082
Binary files /dev/null and b/data/unified_datasets/multiwoz21/data.zip differ
diff --git a/data/unified_datasets/multiwoz21/ontology.json b/data/unified_datasets/multiwoz21/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..b976054f3b6dc47c327c75f545549dfe3563ad01
--- /dev/null
+++ b/data/unified_datasets/multiwoz21/ontology.json
@@ -0,0 +1,1895 @@
+{
+ "domains": {
+ "taxi": {
+ "description": "taxi information query system",
+ "slots": {
+ "destination": {
+ "description": "destination of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_phone": {
+ "description": "taxi phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_types": {
+ "description": "taxi type",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "restaurant": {
+ "description": "restaurant information query system",
+ "slots": {
+ "food": {
+ "description": "food type for the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget for the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "not moderate",
+ "ch",
+ "high - end",
+ "luxury options",
+ "priced cheaply",
+ "moderate",
+ "various",
+ "all price ranges",
+ "cheap price range",
+ "low cost",
+ "assorted",
+ "cheap priced",
+ "moderate or expensive priced",
+ "expensive price",
+ "your",
+ "most expensive",
+ "moderate and expensive",
+ "reasonable",
+ "moderately expensive",
+ "moderate|cheap",
+ "fairly expensive",
+ "upscale",
+ "quite inexpensive",
+ "boderate",
+ "affordable",
+ "expinsive",
+ "they vary",
+ "more moderate",
+ "expensive / upscale",
+ "high class",
+ "moderately price",
+ "reasonably priced",
+ "a little expensive but worth it",
+ "middle",
+ "any price point",
+ "priced moderately",
+ "cherap",
+ "moderately prices",
+ "cheap to expensive",
+ "or otherwise",
+ "inexpensive",
+ "cheaply priced",
+ "adforable",
+ "that range",
+ "great prices",
+ "rather expensive",
+ "expensive side",
+ "moderate range",
+ "cheaply - priced",
+ "more expensive",
+ "varying price",
+ "similar price range",
+ "this price range",
+ "expensive",
+ "expensive or moderate",
+ "not cheap",
+ "the least expensive",
+ "pretty expensive",
+ "high end",
+ "not expensive",
+ "less expensive",
+ "moderatley priced",
+ "most affordable price range",
+ "budget - friendly",
+ "expensive and moderately priced",
+ "moderatre",
+ "cheap range",
+ "or the moderately priced",
+ "fairly cheap",
+ "not too expensive",
+ "east",
+ "fairly inexpensive",
+ "regardless of price",
+ "north",
+ "mostly expensive and moderately priced",
+ "cheap",
+ "budget conscious",
+ "same price range",
+ "very expensive",
+ "not too pricey",
+ "any price range",
+ "quite expensive",
+ "that price range",
+ "your price range",
+ "every price point",
+ "expensively priced",
+ "varying price range",
+ "not - so - expensive",
+ "all",
+ "vietnamese",
+ "any",
+ "moderately - priced",
+ "that price",
+ "moderate pricing",
+ "do nt care",
+ "modest",
+ "pricey",
+ "expensive but worth every penny",
+ "cheaper",
+ "pretty cheap",
+ "moderate price",
+ "ranging from cheap to expensive",
+ "quite low",
+ "moderate priced",
+ "centre",
+ "the same",
+ "expensive range",
+ "relatively cheap",
+ "cheap or expensive",
+ "epensive",
+ "moderate and one in the cheap range",
+ "budget friendly",
+ "fine",
+ "on the cheap side",
+ "expensive price range",
+ "mostly expensive",
+ "moderately priced",
+ "relatively expensive",
+ "moderately",
+ "moderatly",
+ "on the pricey side",
+ "low priced",
+ "expensively",
+ "moderate price range",
+ "moderatly priced",
+ "do n't care"
+ ]
+ },
+ "address": {
+ "description": "exact location of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "near the centre",
+ "east area",
+ "northern parts of town",
+ "any part of town",
+ "west cambridge",
+ "town of centre",
+ "centre of cambridge",
+ "north or centre",
+ "westside",
+ "city centre",
+ "the west end",
+ "the south of town",
+ "the city center",
+ "elsewhere in the city",
+ "town center",
+ "the west side",
+ "south of cambridge",
+ "south area of town",
+ "here",
+ "north side of town",
+ "on the east",
+ "throughout the city",
+ "south area",
+ "centre city",
+ "city",
+ "in the north",
+ "not",
+ "east side",
+ "not in the west",
+ "here in the city",
+ "the south area of town",
+ "that part of town",
+ "the east",
+ "west of town",
+ "centere",
+ "in the city .",
+ "south part of time",
+ "mexican",
+ "the east side",
+ "south part of town",
+ "west part of town .",
+ "southern",
+ "any where in the city",
+ "west area",
+ "north cambridge",
+ "the south area of the city",
+ "almost every area of town",
+ "ely",
+ "the north end",
+ "south cambridge",
+ "the are",
+ "east side of town",
+ "southside",
+ "east side of the city",
+ "by the airport",
+ "around there",
+ "all of cambridge",
+ "the town centre",
+ "nearby",
+ "town centre",
+ "the center part of town",
+ "other parts of town",
+ "not in the south of town",
+ "chesterton",
+ "the centre of town",
+ "east area of town",
+ "the east area",
+ "southend",
+ "ctre",
+ "same area",
+ "anywhere in cambridge",
+ "north side of cambridge",
+ "same area as the park",
+ "south side of town",
+ "the centre area",
+ "the same area as the botanic gardens",
+ "south part of the city",
+ "the area you have chosen",
+ "in the centre",
+ "northern part of cambridge",
+ "east section",
+ "east|south",
+ "the east part of town",
+ "this town",
+ "the same area",
+ "west area of town",
+ "town centre area",
+ "the southern area",
+ "northern part",
+ "southern area",
+ "the east side of town",
+ "expensive",
+ "east section of town",
+ "same",
+ "in cambridge",
+ "north side of chersteron",
+ "towns centre",
+ "in town",
+ "west side",
+ "in the east",
+ "centrally",
+ "west side area",
+ "all over town",
+ "centre region",
+ "anywhere in town",
+ "closely located",
+ "west part of town",
+ "cambridge",
+ "downtown",
+ "south end",
+ "close it city centre",
+ "in the city centre",
+ "close to the hotel",
+ "east",
+ "north and west",
+ "west end of town",
+ "the south part of town",
+ "the north side",
+ "any area",
+ "the area of west",
+ "center of the town",
+ "the area",
+ "the centre part of town",
+ "north",
+ "wet part of town",
+ "other part of the town",
+ "cheap",
+ "north part of town",
+ "centre area of the town",
+ "cetre",
+ "in town at all",
+ "central",
+ "east part of town",
+ "near the center of town",
+ "the city centre",
+ "north end",
+ "east of town",
+ "centreof",
+ "west",
+ "the north side of town",
+ "in the west",
+ "same side of town as your hotel",
+ "the west of town",
+ "centre area of town",
+ "centrem",
+ "east of cambridge",
+ "center of town",
+ "anywhere",
+ "throughout the area",
+ "central area",
+ "south of town",
+ "the west area",
+ "the center of town",
+ "centre part of town",
+ "the center",
+ "north side",
+ "the south",
+ "the west side of town",
+ "west part of the city",
+ "center",
+ "any",
+ "the north",
+ "east end",
+ "the west part of town",
+ "west end",
+ "center cambridge",
+ "the centre area of town",
+ "centre area",
+ "northside",
+ "centre area of the city",
+ "center area of town",
+ "centrally located",
+ "the south side",
+ "around the college",
+ "do nt care",
+ "thai",
+ "the west",
+ "south side",
+ "cambridge centre",
+ "town",
+ "west of cambridge",
+ "near clare hall",
+ "south of town .",
+ "n the centre",
+ "centre of the city",
+ "the north area",
+ "that",
+ "this area",
+ "central region of town",
+ "centre",
+ "north of town",
+ "these areas",
+ "the same",
+ "south part",
+ "that area of town",
+ "in the city",
+ "that area",
+ "city center",
+ "south",
+ "close to your location",
+ "north area of town",
+ "all over cambridge",
+ "the south side of town",
+ "within this area",
+ "near centre",
+ "the centre",
+ "centre of town .",
+ "the north part of town",
+ "the north side .",
+ "same area as the hotel",
+ "and centre",
+ "west side of town",
+ "westies",
+ "around town",
+ "east part of tow",
+ "eastside",
+ "close to that area",
+ "all over the city",
+ "south side of the town",
+ "centre cambridge",
+ "do n't care",
+ "north area",
+ "centre of town"
+ ]
+ },
+ "postcode": {
+ "description": "postcode of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "restaurant phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "time of the restaurant booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the restaurant booking",
+ "is_categorical": true,
+ "possible_values": [
+ "saturday",
+ "wednesday",
+ "tuesday",
+ "w",
+ "saturday|thursday",
+ "monday",
+ "friday",
+ "thursday",
+ "sunday",
+ "sunday|thursday"
+ ]
+ },
+ "choice": {
+ "description": "number of restaurants meeting requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "attraction": {
+ "description": "an entertainment that is offered to the public",
+ "slots": {
+ "address": {
+ "description": "details of where the attraction is",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "phone number of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "entrance fee": {
+ "description": "the fee charged for admission to the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "type of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the attraction",
+ "is_categorical": true,
+ "possible_values": [
+ "east area",
+ "near the centre",
+ "west cambridge",
+ "centre of cambridge",
+ "other areas",
+ "westside",
+ "the west - side",
+ "city centre",
+ "central district",
+ "near the restaurant",
+ "whole of cambridge",
+ "the west end",
+ "entre",
+ "various",
+ "northern area",
+ "town center",
+ "west of your city",
+ "the west side",
+ "south of cambridge",
+ "west part of cambridge",
+ "norwich",
+ "throughout the city",
+ "centre city",
+ "south area",
+ "city",
+ "very close in the same area",
+ "centre by galleria",
+ "east side",
+ "northern",
+ "that part of town",
+ "the east",
+ "same general area",
+ "went side of town",
+ "southern cambridge",
+ "the east side",
+ "south part of town",
+ "in that area",
+ "to the south",
+ "clifton way",
+ "there",
+ "west area",
+ "in the south",
+ "center are",
+ "south cambridge",
+ "ely",
+ "east side of town",
+ "central cambridge",
+ "south park of town",
+ "the south area",
+ "centre area of cambridge",
+ "west area of the city",
+ "several different parts of town",
+ "cambridge 's centre",
+ "nearby",
+ "town centre",
+ "the center part of town",
+ "near the hotel",
+ "museum",
+ "the area you 're looking for",
+ "other parts of town",
+ "most of them are in the center",
+ "the centre of town",
+ "northern area of town",
+ "east area of town",
+ "north of the city",
+ "different area",
+ "same area",
+ "east section",
+ "south side of town",
+ "the centre area",
+ "in the centre",
+ "in the center",
+ "western part of town",
+ "centre area .",
+ "south part of cambridge",
+ "the same road",
+ "cetnre of town",
+ "the same area",
+ "town centre area",
+ "the east part of town",
+ "center area",
+ "west area of town",
+ "cenre",
+ "center of town .",
+ "cambridge leisure park",
+ "church area",
+ "that area .",
+ "near ely",
+ "in that side",
+ "east cambridge",
+ "same",
+ "in cambridge",
+ "towns centre",
+ "that side of town",
+ "another area",
+ "in town",
+ "west side",
+ "west end of the city",
+ "the centry area",
+ "in the east",
+ "we",
+ "all over town",
+ "all around the city",
+ "city centre .",
+ "center of cambridge",
+ "centre region",
+ "west part of town",
+ "cambridge",
+ "cent",
+ "western part of the town",
+ "downtown",
+ "south end",
+ "east",
+ "this side of town",
+ "same area as hotel",
+ "center of the town",
+ "west end of town",
+ "the area",
+ "eat",
+ "northern cambridge",
+ "north",
+ "same part of town as your restaurant",
+ "the centre part of town",
+ "north part of town",
+ "central",
+ "close to the center of town",
+ "east part of town",
+ "same part",
+ "western cambridge",
+ "the city centre",
+ "east of town",
+ "west",
+ "the eastside",
+ "north in milton",
+ "the west of town",
+ "centre area of town",
+ "ce",
+ "east of cambridge",
+ "the east of the town",
+ "close to the restaurant",
+ "the west area of town",
+ "center of town",
+ "the south near your hotel",
+ "south of town",
+ "the west area",
+ "centre part of town",
+ "the center",
+ "north side",
+ "north section of cambridge",
+ "the north of the city",
+ "all of the other areas",
+ "the south",
+ "all",
+ "the west side of town",
+ "west part of the city",
+ "center",
+ "the north",
+ "your desired location",
+ "wet end",
+ "center part of town",
+ "cb30aq",
+ "west end",
+ "in the area",
+ "centre area",
+ "centrally located",
+ "the south side",
+ "city 's centre",
+ "do nt care",
+ "the west",
+ "east end of town",
+ "every area except the north",
+ "south side",
+ "centre of the city",
+ "that",
+ "this area",
+ "centre",
+ "north of town",
+ "right in the center of town",
+ "city cenre",
+ "centre of town ?",
+ "that area",
+ "city center",
+ "south",
+ "same area as tandoori palace",
+ "centre|west",
+ "central zone",
+ "the centre",
+ "the center area",
+ "west side of town",
+ "western area of town",
+ "the center of the park",
+ "west side of the city",
+ "eastside",
+ "northend",
+ "the area you are looking for",
+ "all over the city",
+ "western",
+ "on the centre",
+ "museums",
+ "north area",
+ "centre of town",
+ "center of the city"
+ ]
+ },
+ "name": {
+ "description": "name of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "the price range for the attraction, from cheap to expensive",
+ "is_categorical": true,
+ "possible_values": [
+ "free",
+ "no entrance fee",
+ "entrance fee",
+ "unaware of what their entrance fee is",
+ "not listed",
+ "cheap",
+ "we do n't have any information",
+ "2 pounds",
+ "do n't see a price listed",
+ "from free to 2 pounds 50",
+ "4 pounds",
+ "moderately priced",
+ "neither prices are listed",
+ "expensive",
+ "5 pounds",
+ "they do n't have the entrance fee posted",
+ "free admission",
+ "not sure of the fee",
+ "5 pound entrance fee",
+ "do n't have information",
+ "3.50 pounds"
+ ]
+ },
+ "choice": {
+ "description": "number of attractions matching requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "booking": {
+ "description": "to arrange with a taxi, restaurant, train, etc.",
+ "slots": {
+ "time": {
+ "description": "time for an order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day for an order, from monday to sunday",
+ "is_categorical": true,
+ "possible_values": [
+ "saturday",
+ "wednesday",
+ "saturday night thru monday",
+ "vmhj6y3i",
+ "6",
+ "weds",
+ "tuesday through friday",
+ "tuesday instead of thursday",
+ "2",
+ "that",
+ "tues",
+ "sunday 18/06/2017",
+ "tuesday 's",
+ "today",
+ "tonight",
+ "that time frame",
+ "tuesday",
+ "at that time",
+ "1",
+ "this evening",
+ "sunday and monday",
+ "the same day",
+ "sundar",
+ "monday",
+ "friday",
+ "thursday",
+ "sunday",
+ "we d"
+ ]
+ },
+ "stay": {
+ "description": "for how long the user wish to be at a place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "how many person the order is for",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the ordered place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "train": {
+ "description": "query and order a train",
+ "slots": {
+ "destination": {
+ "description": "destination of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "duration": {
+ "description": "the length of time the train trip lasts",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking for train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the train",
+ "is_categorical": true,
+ "possible_values": [
+ "cambridge",
+ "fr",
+ "saturday",
+ "wednesday",
+ "13:07",
+ "everday",
+ "weds",
+ "sat",
+ "sundays",
+ "train",
+ "that",
+ "between friday and wednesday",
+ "all week",
+ "this day",
+ "every day",
+ "tuesday",
+ "saturdays",
+ "tr2519",
+ "every",
+ "other days are available",
+ "mondays",
+ "we",
+ "monday",
+ "friday",
+ "thursday",
+ "frday",
+ "sunday",
+ "daily",
+ "that day",
+ "fiday"
+ ]
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time for the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "price for the train ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "choice": {
+ "description": "number of trains that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel": {
+ "description": "to query hotel information and place an order",
+ "slots": {
+ "internet": {
+ "description": "internet option at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "free internet",
+ "free",
+ "y",
+ "do nt care",
+ "yes",
+ "none",
+ "no",
+ "4"
+ ]
+ },
+ "area": {
+ "description": "area or place of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "near the centre",
+ "any part of town",
+ "east area",
+ "different parts of the city",
+ "north end of the city",
+ "centre of cambridge",
+ "near the restaurant",
+ "city centre",
+ "throughout cambridge",
+ "different part of town",
+ "eastern part of the city",
+ "the west end",
+ "various",
+ "northern area",
+ "the northern part of town",
+ "northern part of town",
+ "the west side",
+ "on the west",
+ "south area of town",
+ "centrally - located",
+ "north side of town",
+ "central location",
+ "centra",
+ "throughout the city",
+ "not in the south",
+ "south area",
+ "centre city",
+ "in the north",
+ "east side",
+ "northern side of town",
+ "different parts of town",
+ "you are interested in",
+ "northern",
+ "that part of town",
+ "west of town",
+ "the east",
+ "all areas",
+ "all parts of the city",
+ "the east side",
+ "south part of town",
+ "south end of the city",
+ "there",
+ "west area",
+ "north cambridge",
+ "the north end",
+ "south cambridge",
+ "east side of town",
+ "southside",
+ "the south area",
+ "or west ?",
+ "another area of town",
+ "all of cambridge",
+ "the town centre",
+ "east near other shops and boutiques",
+ "town centre",
+ "on the west side",
+ "all over",
+ "the centre of the city",
+ "other parts of town",
+ "the centre of town",
+ "east area of town",
+ "the east area",
+ "the westside",
+ "southend",
+ "west|centre",
+ "northe part of town",
+ "anywhere in cambridge",
+ "different area",
+ "same area",
+ "south side of town",
+ "the centre area",
+ "on the south",
+ "in the centre",
+ "in the center",
+ "wast",
+ "all across town",
+ "south end of town",
+ "except in the north",
+ "and east",
+ "place to be a guesthouse",
+ "the same area",
+ "other parts of the city",
+ "the east side of town",
+ "somewhere else",
+ "and in the centre",
+ "east section of town",
+ "east cambridge",
+ "across cambridge",
+ "north of cambridge",
+ "next door",
+ "except in the east",
+ "in cambridge",
+ "north part of the city",
+ "that side of town",
+ "south cambridge area",
+ "in town",
+ "west side",
+ "west end of the city",
+ "we",
+ "centrally",
+ "west side near the restaurant",
+ "all over town",
+ "southern part of town",
+ "all around the city",
+ "west part of town",
+ "cambridge",
+ "downtown",
+ "south end",
+ "east",
+ "east and the north .",
+ "norht",
+ "the north side",
+ "any area",
+ "center of the town",
+ "the area",
+ "eat",
+ "northern cambridge",
+ "north",
+ "cheap",
+ "north part of town",
+ "centre part",
+ "east part of town",
+ "western cambridge",
+ "north end of town",
+ "north end",
+ "east of town",
+ "several areas of town",
+ "west",
+ "eastern",
+ "in the west",
+ "everywhere but the city centre",
+ "centre area of town",
+ "center of town",
+ "west part",
+ "the town center",
+ "south of town",
+ "the center of town",
+ "the west area",
+ "centre part of town",
+ "nborth",
+ "north side",
+ "the north of the city",
+ "the south",
+ "west part of the city",
+ "center",
+ "the north",
+ "west end",
+ "same side of town",
+ "in the area",
+ "any part of the city",
+ "centre area",
+ "not to far from the restaurant",
+ "northside",
+ "on the eastside",
+ "close to where you 'll be dining",
+ "the east end of town",
+ "various parts of the city",
+ "west areas of town",
+ "centrally located",
+ "the south side",
+ "north part of town .",
+ "do nt care",
+ "the west",
+ "same area as the restaurant",
+ "that vicinity",
+ "nearby the restaurant",
+ "south side",
+ "east part of time",
+ "centre of the city",
+ "south closer to the museum",
+ "north park of town",
+ "centre",
+ "north of town",
+ "north and centre",
+ "the same",
+ "on the north",
+ "that area of town",
+ "that area",
+ "in the city",
+ "el shaddai",
+ "city center",
+ "south",
+ "the city 's south side",
+ "north area of town",
+ "near the museum",
+ "eastern cambridge",
+ "near centre",
+ "the easy",
+ "a different area",
+ "the centre",
+ "the north part of town",
+ "west side of town",
+ "across town",
+ "eastside",
+ "northend",
+ "north par of town",
+ "the centre of cambridge",
+ "all over the city",
+ "north location",
+ "centre cambridge",
+ "that region",
+ "north area",
+ "centre of town"
+ ]
+ },
+ "stars": {
+ "description": "star rating of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "4|5",
+ "several",
+ "0 to 4",
+ "no star rating",
+ "do nt care",
+ "one",
+ "does not show",
+ "foru",
+ "does not have",
+ "different star ratings",
+ "unrated",
+ "2",
+ "four starts",
+ "4-star",
+ "four - star",
+ "3|4",
+ "yes",
+ "three",
+ "four",
+ "lower",
+ "four stars",
+ "zero",
+ "ranging from 2 - 4 stars",
+ "two",
+ "1",
+ "3",
+ "four star",
+ "0-star",
+ "five",
+ "drop the star rating",
+ "2-star",
+ "no",
+ "3-star",
+ "5-star",
+ "0",
+ "1-star",
+ "not rated",
+ "5",
+ "not as fancy",
+ "4"
+ ]
+ },
+ "parking": {
+ "description": "parking facility at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "n",
+ "free",
+ "do nt care",
+ "yes",
+ "none",
+ "no",
+ "free parking"
+ ]
+ },
+ "phone": {
+ "description": "hotel phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "cheapest",
+ "economically priced",
+ "free",
+ "all different price ranges",
+ "more moderately priced",
+ "expensively - priced",
+ "ca n't view that information",
+ "moderate",
+ "super cheap",
+ "low cost",
+ "cheap price range",
+ "lower",
+ "affordable pricing",
+ "moderate to cheap",
+ "cheap priced",
+ "inexpensively - priced",
+ "mostly in the cheap to moderate price range",
+ "moderately pricing",
+ "guesthouses",
+ "moderate|cheap",
+ "moderately priceed",
+ "very cheap",
+ "even cheaper",
+ "upscale",
+ "other ranges",
+ "moderate or cheap",
+ "cheaply",
+ "affordable",
+ "different price range",
+ "does not say",
+ "moderately price",
+ "expensive or cheap",
+ "priced moderately",
+ "moderatly price",
+ "moderate price point",
+ "moderately prices",
+ "cheap to expensive",
+ "on the more expensive side",
+ "inexpensive",
+ "the cheapest",
+ "cheaply priced",
+ "that range",
+ "rather expensive",
+ "your price",
+ "moderate range",
+ "cheaply - priced",
+ "more expensive",
+ "expensive",
+ "much cheaper",
+ "cheap side",
+ "slightly more expensive",
+ "$100",
+ "cheaper than the others",
+ "pretty expensive",
+ "not expensive",
+ "cheaply prices",
+ "moderatley priced",
+ "modrate",
+ "cheap range",
+ "espensive",
+ "chear",
+ "fairly cheap",
+ "oderately priced",
+ "moderate in price",
+ "different",
+ "very affordable",
+ "your chosen",
+ "unfortunately do not have the price",
+ "north",
+ "less costly",
+ "different price ranges",
+ "cheap",
+ "very inexpensive",
+ "2",
+ "moderately priced .",
+ "moderately to expensively priced",
+ "economical",
+ "same price range",
+ "moderate to cheap range",
+ "cheap to moderate",
+ "quite expensive",
+ "that price range",
+ "cheap|moderate",
+ "your price range",
+ "varying price ranges",
+ "expensive to moderate",
+ "expensively priced",
+ "epxensive",
+ "or expensive ?",
+ "all",
+ "higher price range",
+ "any",
+ "moderately - priced",
+ "hotel",
+ "moderate pricing",
+ "do nt care",
+ "lower end",
+ "quite cheap",
+ "more budget - friendly",
+ "cheaper",
+ "fairly cheap compared to other hotels",
+ "pretty cheap",
+ "moderate price",
+ "that",
+ "moderate priced",
+ "centre",
+ "the same",
+ "expensive range",
+ "relatively cheap",
+ "moderate prices",
+ "hotels",
+ "cheap>moderate",
+ "a little pricey",
+ "on the cheap side",
+ "moderately - priced or cheap",
+ "expensive price range",
+ "moderately priced",
+ "same price",
+ "moderately",
+ "budget - priced",
+ "moderately pried",
+ "moderate price range",
+ "great",
+ "moderatly priced",
+ "extremely reasonable",
+ "cheap or moderate",
+ "cheaper side"
+ ]
+ },
+ "people": {
+ "description": "number of people for the hotel booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "exact location of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "what is the type of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "stay": {
+ "description": "length of stay at the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the hotel booking",
+ "is_categorical": true,
+ "possible_values": [
+ "next friday",
+ "saturday",
+ "wednesday",
+ "tuesday",
+ "saturday|tuesday",
+ "t",
+ "sunday>monday",
+ "friday>tuesday",
+ "monday",
+ "friday",
+ "thursday",
+ "monday<thursday",
+ "sunday",
+ "wednesday|friday",
+ "monda"
+ ]
+ },
+ "choice": {
+ "description": "number of hotels that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "police": {
+ "description": "find police stations",
+ "slots": {
+ "address": {
+ "description": "exact location of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "police station phone number",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "inform user of value for a certain slot"
+ },
+ "request": {
+ "description": "ask for value of a slot"
+ },
+ "nobook": {
+ "description": "inform user of booking failure"
+ },
+ "reqmore": {
+ "description": "ask user for more instructions"
+ },
+ "book": {
+ "description": "place an order for user"
+ },
+ "bye": {
+ "description": "end a conversation and say goodbye to user"
+ },
+ "thank": {
+ "description": "express gratitude"
+ },
+ "welcome": {
+ "description": "welcome"
+ },
+ "recommend": {
+ "description": "recommend a choice for user request"
+ },
+ "offerbook": {
+ "description": "offer to place an order for user"
+ },
+ "offerbooked": {
+ "description": "inform user that an order is succussful"
+ },
+ "greet": {
+ "description": "express greeting"
+ },
+ "nooffer": {
+ "description": "inform user that no options matches user request"
+ },
+ "select": {
+ "description": "provide several choices for user to choose from"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "stay",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "reqmore",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "bye",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "thank",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "welcome",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "offerbook",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "greet",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "internet",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "parking",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "food",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "nobook",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "stars",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "entrance fee",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "Ref",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "book",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "taxi",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_types",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "offerbooked",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "state": {
+ "hotel": {
+ "name": "",
+ "area": "",
+ "parking": "",
+ "pricerange": "",
+ "stars": "",
+ "internet": "",
+ "type": "",
+ "stay": "",
+ "day": "",
+ "people": ""
+ },
+ "train": {
+ "leaveAt": "",
+ "destination": "",
+ "day": "",
+ "arriveBy": "",
+ "departure": "",
+ "people": ""
+ },
+ "taxi": {
+ "leaveAt": "",
+ "destination": "",
+ "departure": "",
+ "arriveBy": ""
+ },
+ "restaurant": {
+ "food": "",
+ "pricerange": "",
+ "name": "",
+ "area": "",
+ "time": "",
+ "day": "",
+ "people": ""
+ },
+ "attraction": {
+ "type": "",
+ "name": "",
+ "area": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/multiwoz21/original_data.zip b/data/unified_datasets/multiwoz21/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..b27361772be980013a94c91898fb61e3e122ba8e
Binary files /dev/null and b/data/unified_datasets/multiwoz21/original_data.zip differ
diff --git a/data/unified_datasets/multiwoz21/preprocess.py b/data/unified_datasets/multiwoz21/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..5ca81e2e399d8a816e3be6d4ae33babfc2f1597e
--- /dev/null
+++ b/data/unified_datasets/multiwoz21/preprocess.py
@@ -0,0 +1,1431 @@
+import copy
+import re
+import zipfile
+import json
+import os
+from tqdm import tqdm
+import sys
+import difflib
+from fuzzywuzzy import fuzz
+sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
+
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+import logging
+
+
+logging.basicConfig(level=logging.INFO)
+self_dir = (os.path.abspath(os.getcwd()))
+
+REF_SYS_DA = {
+ 'Attraction': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Fee': "entrance fee", 'Name': "name", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange", 'Type': "type",
+ 'none': None, 'Open': None
+ },
+ 'Hospital': {
+ 'Department': 'department', 'Addr': 'address', 'Post': 'postcode',
+ 'Phone': 'phone', 'none': None
+ },
+ 'Booking': {
+ 'Day': 'day', 'Name': 'name', 'People': 'people',
+ 'Ref': 'Ref', 'Stay': 'stay', 'Time': 'time',
+ 'none': None
+ },
+ 'Hotel': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Internet': "internet", 'Name': "name", 'Parking': "parking",
+ 'Phone': "phone", 'Post': "postcode", 'Price': "pricerange",
+ 'Stars': "stars", 'Type': "type", 'Stay': 'stay', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Restaurant': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Name': "name", 'Food': "food", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange",
+ 'Time': 'time', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Taxi': {
+ 'Arrive': "arriveBy", 'Car': "taxi_types", 'Depart': "departure",
+ 'Dest': "destination", 'Leave': "leaveAt", 'Phone': "taxi_phone",
+ 'none': None
+ },
+ 'Train': {
+ 'Arrive': "arriveBy", 'Choice': "choice", 'Day': "day",
+ 'Depart': "departure", 'Dest': "destination",
+ 'Leave': "leaveAt", 'People': "people", 'Ref': "Ref",
+ 'Time': "duration", 'none': None, 'Ticket': 'price',
+ },
+ 'Police': {
+ 'Addr': "address", 'Post': "postcode", 'Phone': "phone", 'none': None
+ },
+}
+
+# taxi restaurant attraction train
+slot_to_type = {
+ 'taxi-destination': 'non',
+ 'taxi-departure': 'non',
+ 'taxi-leaveAt': 'non',
+ 'taxi-arriveBy': 'non',
+ 'restaurant-food': 'non',
+ 'restaurant-name': 'non',
+ 'attraction-address': 'non',
+ 'attraction-postcode': 'non',
+ 'restaurant-pricerange': 'cat',
+ 'restaurant-address': 'non',
+ 'restaurant-area': 'cat',
+ 'restaurant-postcode': 'non',
+ 'attraction-phone': 'non',
+ 'attraction-entrance fee': 'non',
+ 'booking-time': 'non',
+ 'booking-day': 'cat',
+ 'attraction-type': 'non',
+ 'attraction-area': 'cat',
+ 'train-destination': 'non',
+ 'train-arriveBy': 'non',
+ 'train-departure': 'non',
+ 'hotel-internet': 'cat',
+ 'hotel-area': 'cat',
+ 'booking-stay': 'non',
+ 'booking-people': 'non',
+ 'train-duration': 'non',
+ 'train-people': 'non',
+ 'train-day': 'cat',
+ 'train-Ref': 'non',
+ 'hotel-stars': 'cat',
+ 'train-leaveAt': 'non',
+ 'train-price': 'non',
+ 'hotel-parking': 'cat',
+ 'hotel-phone': 'non',
+ 'hotel-name': 'non',
+ 'hotel-pricerange': 'cat',
+ 'hotel-people': 'non',
+ 'restaurant-phone': 'non',
+ 'hotel-postcode': 'non',
+ 'hotel-address': 'non',
+ 'attraction-name': 'non',
+ 'hotel-type': 'non',
+ 'restaurant-people': 'non',
+ 'train-choice': 'non',
+ 'attraction-pricerange': 'cat',
+ 'hotel-stay': 'non',
+ 'booking-name': 'non',
+ 'booking-Ref': 'non',
+ 'restaurant-time': 'non',
+ 'restaurant-day': 'cat',
+ 'hotel-day': 'cat',
+ 'hotel-choice': 'non',
+ 'restaurant-choice': 'non',
+ 'attraction-choice': 'non',
+ 'taxi-taxi_phone': 'non',
+ 'taxi-taxi_types': 'non',
+ 'police-address': 'non',
+ 'police-postcode': 'non',
+ 'police-phone': 'non'
+}
+
+state_cat_slot_value_dict = {
+ "hotel-pricerange": {
+ "cheap": 735,
+ "moderate": 1063,
+ "expensive": 594,
+ },
+ "hotel-parking": {
+ "yes": 1809,
+ "no": 126,
+ "free": 4,
+ },
+ "hotel-day": {
+ "tuesday": 385,
+ "wednesday": 410,
+ "monday": 365,
+ "saturday": 407,
+ "friday": 393,
+ "thursday": 384,
+ "sunday": 369,
+ },
+ "train-day": {
+ "wednesday": 533,
+ "monday": 533,
+ "saturday": 543,
+ "thursday": 547,
+ "friday": 563,
+ "tuesday": 553,
+ "sunday": 613,
+ },
+ "hotel-stars": {
+ "4": 1263,
+ "2": 193,
+ "0": 201,
+ "3": 401,
+ "5": 45,
+ "1": 45,
+ },
+ "hotel-internet": {
+ "yes": 1841,
+ "no": 79,
+ "free": 2
+ },
+ "hotel-area": {
+ "east": 416,
+ "north": 717,
+ "centre": 538,
+ "south": 289,
+ "west": 316,
+ },
+ "attraction-area": {
+ "centre": 1290,
+ "west": 332,
+ "north": 155,
+ "south": 240,
+ "east": 272,
+ },
+ "restaurant-pricerange": {
+ "expensive": 1477,
+ "cheap": 758,
+ "moderate": 1028,
+ },
+ "restaurant-area": {
+ "centre": 1745,
+ "south": 398,
+ "north": 390,
+ "east": 360,
+ "west": 423,
+ },
+ "restaurant-day": {
+ "thursday": 362,
+ "wednesday": 412,
+ "friday": 395,
+ "monday": 383,
+ "sunday": 399,
+ "saturday": 421,
+ "tuesday": 350,
+ }
+}
+
+
+synonyms = [
+ ["el shaddia guesthouse", "el shaddai"],
+ [ "peterborough", "peterbourgh"],
+ ["night club", "nightclub", 'nightclubs'],
+ ["boat", "boating"],
+ ["portugese", "portuguese"],
+ ["guesthouse", "guest house"],
+ ["seafood", "sea food"],
+ ["christ 's college", "christ college"],
+ ["huntingdon marriott hotel"]
+]
+
+state_cat_slot_ds = [k for k, v in slot_to_type.items() if v == 'cat']
+
+da_cat_slot_values = {
+ # 'hotel-stay': ['1', '2', '3', '4', '5'],
+ 'hotel-internet': ['free', 'no', 'none', 'yes'],
+ 'hotel-parking': ['free', 'no', 'none', 'yes']
+}
+
+state_cat_slot_values = {}
+
+multiwoz_desc = {
+ 'taxi': {
+ 'domain': 'taxi information query system',
+ 'taxi_phone': 'taxi phone number',
+ 'taxi_types': 'taxi type',
+ },
+ 'restaurant': {
+ 'domain': 'restaurant information query system',
+ 'address': 'exact location of the restaurant',
+ 'postcode': 'postcode of the restaurant',
+ 'phone': 'restaurant phone number',
+ 'choice': 'number of restaurants meeting requests of user',
+ },
+ 'attraction': {
+ 'domain': 'an entertainment that is offered to the public',
+ 'address': 'details of where the attraction is',
+ 'postcode': 'postcode of the attraction',
+ 'phone': 'phone number of the attraction',
+ 'entrance fee': 'the fee charged for admission to the attraction',
+ 'pricerange': 'the price range for the attraction, from cheap to expensive',
+ 'choice': 'number of attractions matching requests of user'
+ },
+ 'booking': {
+ 'domain': 'to arrange with a taxi, restaurant, train, etc.',
+ 'time': 'time for an order',
+ 'day': 'day for an order, from monday to sunday',
+ 'stay': 'for how long the user wish to be at a place',
+ 'people': 'how many person the order is for',
+ 'name': 'name of the ordered place',
+ 'Ref': 'reference number of the order'
+ },
+ 'train': {
+ 'domain': 'query and order a train',
+ 'duration': 'the length of time the train trip lasts',
+ 'Ref': 'reference number of the order',
+ 'price': 'price for the train ticket',
+ 'choice': 'number of trains that meets requests of the user',
+ },
+ 'hotel': {
+ 'domain': 'to query hotel information and place an order',
+ 'address': 'exact location of the hotel',
+ 'postcode': 'postcode of the hotel',
+ 'phone': 'hotel phone number',
+ 'choice': 'number of hotels that meets requests of the user',
+ },
+ 'police': {
+ 'domain': 'find police stations',
+ 'address': 'exact location of the police station',
+ 'postcode': 'postcode of the police station',
+ 'phone': 'police station phone number',
+ },
+ 'intents': {
+ 'inform': 'inform user of value for a certain slot',
+ 'request': 'ask for value of a slot',
+ 'nobook': 'inform user of booking failure',
+ 'reqmore': 'ask user for more instructions',
+ 'book': 'place an order for user',
+ 'bye': 'end a conversation and say goodbye to user',
+ 'thank': 'express gratitude',
+ 'welcome': 'welcome',
+ 'offerbooked': 'inform user that an order is succussful',
+ 'recommend': 'recommend a choice for user request',
+ 'greet': 'express greeting',
+ 'nooffer': 'inform user that no options matches user request',
+ 'offerbook': 'offer to place an order for user',
+ 'select': 'provide several choices for user to choose from',
+ }
+}
+
+digit2word = {
+ '0': 'zero', '1': 'one', '2': 'two', '3': 'three', '4': 'four', '5': 'five',
+ '6': 'six', '7': 'seven', '8': 'eight', '9': 'nine', '10': 'ten'
+}
+
+
+def pharse_in_sen(phrase, sen):
+ '''
+ match value in the sentence
+ :param phrase: str
+ :param sen: str
+ :return: start, end if matched, else None, None
+ '''
+ assert isinstance(phrase, str)
+ pw = '(^|[\s,\.:\?!-])(?P<v>{})([\s,\.:\?!-]|$)'
+ pn = '(^|[\s\?!-]|\D[,\.:])(?P<v>{})($|[\s\?!-]|[,\.:]\D|[,\.:]$)'
+
+ if phrase.isdigit() and phrase in digit2word:
+ phrase = digit2word[phrase]
+ p = re.compile(pw.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ # match['0'] += 1
+ if phrase.isdigit():
+ pattern = pn
+ else:
+ pattern = pw
+ p = re.compile(pattern.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ return (None, None), 0
+
+
+
+
+
+def update_state(state, update):
+ # print('======================')
+ # print(state)
+ # print(update)
+ # print('======================')
+
+ for service, service_update in update.items():
+ if service not in state:
+ state[service] = copy.deepcopy(service_update)
+ else:
+ state[service].update(update[service])
+
+
+def convert_da(utt, da_dict, binary_ont, intent_ont, did, tid, da_cat_slot_values):
+ '''
+ convert multiwoz dialogue acts to required format
+ :param utt: user or system utt
+ :param da_dict: multiwoz da
+ :param binary_ont: binary ontology
+ :param intent_ont: intent ontology
+ :return:
+ '''
+ converted_da = {
+ 'categorical': [],
+ 'non-categorical': [],
+ 'binary': []
+ }
+
+ for Domain_Act, S, v in da_dict:
+ Domain, Act = Domain_Act.split('-')
+ if Domain.lower() in ['police', 'hospital', 'bus']:
+ continue
+
+ if Act.lower() not in intent_ont:
+ intent_ont[Act.lower()] = {}
+
+ # general domain is converted to empty domain. e.g. thank, bye
+ if Domain == 'general':
+ assert S == 'none'
+ assert v == 'none'
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': '',
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+
+
+ try:
+ reformated_slot = REF_SYS_DA[Domain][S]
+ except:
+ # print('44444444444444444444444444444444')
+ # print(Domain, S)
+ # logging.info('slot not in REF_SYS_DA, drop')
+ continue
+
+ # if slot is None, da should be converted into binary
+ if reformated_slot is None:
+ if not (S == 'none' and v == 'none'):
+ # mainly for `Open` slot
+ # print('11111111111111111111')
+ # print(Domain_Act, S, v)
+ continue
+ # Booking-Inform none none
+ # Police-Inform none none
+ # Train-OfferBook none none
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ reformated_domain_slot = Domain.lower() + '-' + reformated_slot
+
+ if Act.lower() == 'request':
+ converted_dict = {
+ 'intent': 'request',
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ # vs = da_dict[(Domain_Act, S)]['values']
+
+ if reformated_domain_slot in slot_to_type and slot_to_type[reformated_domain_slot] == 'cat':
+ origin_v = v
+ v = v.lower()
+ # if reformated_domain_slot in cat_slot_proj:
+ # v = cat_slot_proj[reformated_domain_slot][v]
+ if reformated_domain_slot not in da_cat_slot_values:
+ da_cat_slot_values[reformated_domain_slot] = []
+ # if v not in cat_slot_values[reformated_domain_slot]:
+ da_cat_slot_values[reformated_domain_slot].append(v)
+ converted_da['categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ if 'start_word' in da_dict[(Domain_Act, S, origin_v)]:
+ start_ws = da_dict[(Domain_Act, S, origin_v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, origin_v)]['end_word']
+ utt_list = utt.split()
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w)
+ if not utt[start_ch: end_ch] == origin_v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ converted_da['categorical'][-1].update({
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break
+
+ else:
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ # todo no span annotation
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ continue
+
+ start_ws = da_dict[(Domain_Act, S, v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, v)]['end_word']
+ utt_list = utt.split()
+ found = True
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w, v)
+ if not utt[start_ch: end_ch] == v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ found = True
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break
+
+ if not found:
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ return converted_da
+
+
+def get_state_update(prev_state, cur_state, dialog, did, tid, utt, coref_dict, slot_notfound_dict, da_cat_slot_values):
+ prev_turns = dialog['turns']
+ state_update = {'categorical': [], 'non-categorical': []}
+ notfoundnum = 0
+ total_value = 0
+
+ diff_state = {}
+ if prev_state is None:
+ diff_state = {domain: {slot: value for slot, value in cur_state[domain].items() if value != ''} for domain in
+ cur_state}
+ else:
+ assert len(prev_state) == len(cur_state), print(prev_state, cur_state)
+ for domain, domain_state in prev_state.items():
+ if domain not in diff_state:
+ diff_state[domain] = {}
+ for slot, value in domain_state.items():
+ if value != cur_state[domain][slot]:
+ # assert len(cur_state[domain][slot]) > 0, print(did, tid, domain, slot, utt)
+ diff_state[domain][slot] = cur_state[domain][slot]
+
+ ret_diff_state = copy.deepcopy(diff_state)
+
+
+
+ for domain in diff_state:
+ for slot in diff_state[domain]:
+
+ total_value += 1
+ fix_or = False
+ if '|' in diff_state[domain][slot]:
+ value = diff_state[domain][slot].split('|')[0]
+ else:
+ value = diff_state[domain][slot]
+
+ # if dialog['original_id'] == 'PMUL2512' and tid == 17 and value == '02:45':
+ # value = '2:45'
+
+ value_list = [value]
+ for _synonyms in synonyms:
+ if value in _synonyms:
+ value_list = _synonyms
+
+ value_list.extend(get_time_variants(value))
+ value_list.extend(get_genitive_variants(value))
+ value_list.extend(get_bb_variants(value))
+
+ if value.endswith(' restaurant'):
+ value_list.append(value.split(' restaurant')[0])
+ if value.endswith(' hotel'):
+ value_list.append(value.split(' hotel')[0])
+ found = False
+ for value in value_list:
+ # categorical slots
+ if slot in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']:
+ reformated_domain_slot = '-'.join([domain, slot])
+ if reformated_domain_slot in state_cat_slot_value_dict and (value in state_cat_slot_value_dict[reformated_domain_slot] or value in ['dontcare', '', 'none', 'not mentioned']):
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = [diff_state[domain][slot]]
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ else :
+ for _turn in prev_turns[::-1]:
+ found = False
+ for da in _turn['dialogue_act']['categorical']:
+ if da['value'] == value:
+ if 'start' in da:
+ state_update['categorical'][-1].update({
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ found = True
+ break
+ if found:
+ break
+ else:
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = [diff_state[domain][slot]]
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ # reformated_domain_slot = '-'.join([domain, slot]
+ found = True
+ break
+
+ # process value ---> none
+ assert value not in ['none', 'not mentioned']
+ if value in ['', 'dontcare']:
+ # if reformated_domain_slot not in state_cat_slot_values:
+ # state_cat_slot_values[reformated_domain_slot] = []
+ # # if v not in cat_slot_values[reformated_domain_slot]:
+ # state_cat_slot_values[reformated_domain_slot].append(value)
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ found = True
+ break
+
+ # first look for values in coref_dict
+ for _Domain_Act, _Slot, _value in coref_dict:
+ _domain, _act = _Domain_Act.lower().split('-')
+ _slot = _Slot.lower()
+ _coref_value = coref_dict[(_Domain_Act, _Slot, _value)]['coref_value']
+ if _coref_value == '':
+ continue
+ _coref_turn = coref_dict[(_Domain_Act, _Slot, _value)]['turn']
+ if _coref_turn == -1:
+ continue
+ _coref_pos = coref_dict[(_Domain_Act, _Slot, _value)]['pos']
+ if _coref_pos == '':
+ continue
+ _utt = coref_dict[(_Domain_Act, _Slot, _value)]['utt']
+ if _domain == domain and _slot == slot and value == _coref_value:
+
+ start_w, end_w = [int(p) for p in _coref_pos.split('-')]
+ utt_list = _utt.split()
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w + 1):
+ end_ch += len(utt_list[i]) + 1
+ end_ch -= 1
+
+ if not _utt[start_ch: end_ch] == _coref_value:
+ # print(111111111111111111111111111111111)
+ # print(_utt[start_ch: end_ch], _coref_value)
+ continue
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'from': 'coref',
+ 'utt_idx': _coref_turn,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ found = True
+
+ if found:
+ break
+
+ # from da annotation
+ for _turn in prev_turns[::-1]:
+ for da in _turn['dialogue_act']['non-categorical']:
+ # if da['domain'] == domain and da['slot'] == slot and fuzz.ratio(da['value'], value) > 85:
+ # if not da['value'] == value:
+ # print(1111111111111111)
+ # print(value, da['value'])
+
+ if fuzz.ratio(da['value'], value) > 85:
+
+ if 'start' in da:
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ # 'value': da['value'],
+ 'value': diff_state[domain][slot],
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if da['value'] != value:
+ state_update['non-categorical'][-1].update({'fixed_value':da['value']})
+ ret_diff_state[domain][slot] = da['value']
+
+ break
+ if found:
+ break
+
+ if found:
+ break
+
+ # from utterance
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+ (start, end), num = pharse_in_sen(str(value), _utt)
+ if num:
+ assert value.lower() == _utt[start:end].lower() \
+ or digit2word[value].lower() == _utt[start:end].lower()
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': _utt[start:end].lower(),
+ # 'fixed_value': _utt[start:end].lower(),
+ 'from': 'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': start,
+ 'end': end
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if value != _utt[start:end].lower():
+ state_update['non-categorical'][-1].update({'fixed_value': _utt[start:end].lower()})
+ ret_diff_state[domain][slot] = _utt[start:end].lower()
+ found = True
+ break
+ if found:
+ break
+
+ # from utterance
+ if not value.isdigit():
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+
+ s = difflib.SequenceMatcher(None, _utt, value)
+ matches = s.get_matching_blocks()
+
+ for i, j, n in matches:
+ possible_value = _utt[i: i+len(value)]
+
+ if i+ len(value) < len(_utt) and _utt[i+len(value)] not in [ ' ', ',', '.', '?', '!', '/'] :
+ possible_value += _utt[i+len(value):].split()[0]
+
+ if possible_value.startswith('th '):
+ possible_value = possible_value[3:]
+ i += 3
+ if i > 0 and _utt[i-1] not in [ ' ', ',', '.', '?', '!', '/']:
+ # cut first incomplete word
+ if len(possible_value.split()) > 1:
+ i += len(possible_value.split()[0]) + 1
+ possible_value = ' '.join(possible_value.split()[1:])
+
+
+ # prepend first incomplete word
+ # possible_value = _utt[:i].split()[-1] + possible_value
+ # i -= len(_utt[:i].split()[-1])
+
+
+ if fuzz.token_sort_ratio(value, possible_value) > 92 or possible_value.startswith('ashley hotel and lovell lodge') :
+ found = True
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': possible_value,
+ # 'fixed_value': possible_value,
+ 'from':'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': i,
+ 'end': i+len(possible_value)
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if possible_value != value:
+ state_update['non-categorical'][-1].update({'fixed_value': possible_value})
+ ret_diff_state[domain][slot] = possible_value
+ break
+ # assert _utt[i:i+len(possible_value)] == possible_value, print(_utt, _utt[i:i+len(possible_value)], possible_value)
+ # break
+ # if not possible_value == value:
+ # print(3333333333333333)
+ # print(value)
+ # print(possible_value)
+ if found:
+ break
+ if found:
+ break
+
+ if found:
+ break
+ if not found:
+ # print('3333333333333333333')
+ # print(did, tid)
+ # print(domain, slot, value)
+ # print([_t['utterance'] for _t in prev_turns])
+ # assert slot not in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']
+
+ if (domain, slot) not in slot_notfound_dict:
+ slot_notfound_dict[(domain, slot)] = 1
+ else:
+ slot_notfound_dict[(domain, slot)] += 1
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ return state_update, notfoundnum, total_value, ret_diff_state
+
+
+def merge_data_annotation():
+ extract_dir = os.path.join(self_dir, 'original_data')
+ data21 = json.load(open(os.path.join(self_dir, extract_dir, 'data.json')))
+ # data21_train = json.load(open(os.path.join(self_dir, extract_dir, 'train.json')))
+ # data21_val = json.load(open(os.path.join(self_dir, extract_dir, 'val.json')))
+ # data21_test = json.load(open(os.path.join(self_dir, extract_dir, 'test.json')))
+ # data21 = {}
+ # data21.update(data21_train)
+ # data21.update(data21_val)
+ # data21.update(data21_test)
+
+ # update_from_25_cnt = 0
+ # total_turn = 0
+ # for dial_id, dialog in data21.items():
+ # dial_id = dial_id + '.json'
+ # assert dial_id in data25
+ # for i, _turn in enumerate(dialog['log']):
+ # total_turn += 1
+ # if _turn['text'] == data25[dial_id]['log'][i]['text']:
+ # _turn['span_info'].extend(copy.deepcopy(data25[dial_id]['log'][i]['span_info']))
+ # # _turn['span_info'] = list(set(_turn['span_info']))
+ # # _turn['dialog_act'].update(copy.deepcopy(data25[dial_id]['log'][i]['dialog_act']))
+ # for Domain_Intent in data25[dial_id]['log'][i]['dialog_act']:
+ # if Domain_Intent in _turn['dialog_act']:
+ # _turn['dialog_act'][Domain_Intent].extend(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # else:
+ # _turn['dialog_act'][Domain_Intent] = copy.deepcopy(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # # _turn['dialog_act'][Domain_Intent] = list(set(_turn['dialog_act'][Domain_Intent]))
+ # if 'coreference' in data25[dial_id]['log'][i]:
+ # _turn['coreference'] = copy.deepcopy(data25[dial_id]['log'][i]['coreference'])
+ # update_from_25_cnt += 1
+ # else:
+ # # print('==============multiwoz21=================')
+ # # print(_turn['text'])
+ # # print('==============multiwoz25=================')
+ # # print(data25[dial_id]['log'][i]['text'])
+ # continue
+ #
+ # print('{}/{} turns update from multiwoz25 data'.format(update_from_25_cnt, total_turn))
+ return data21
+
+
+def preprocess(da_cat_slot_values, state_cat_slot_values):
+ all_data = []
+ binary_ont = []
+ intent_ont = {}
+ state_ont = {}
+
+ data_splits = ['train', 'val', 'test']
+ # data_splits = ['test']
+ extract_dir = os.path.join(self_dir, 'original_data')
+ num_train_dialogue = 0
+ num_train_utt = 0
+
+ num_match_error_da_span = 0
+
+ if not os.path.exists('data.zip') or not os.path.exists('ontology.json'):
+ # for data_split in data_splits:
+ # data_zip_file = os.path.join(self_dir, 'original_data.zip')
+ # if not os.path.exists(data_zip_file):
+ # raise FileNotFoundError(data_zip_file)
+
+ # logging.info('unzip multiwoz data to {}'.format(extract_dir))
+ # archive = zipfile.ZipFile(data_zip_file, 'r')
+ # archive.extractall(extract_dir)
+
+ data = merge_data_annotation()
+ # exit()
+ # data = json.load(open(os.path.join(self_dir, extract_dir, 'data_meta_fixed.json')))
+ train_list = open(os.path.join(self_dir, extract_dir, 'trainListFile')).read().split()
+ val_list = open(os.path.join(self_dir, extract_dir, 'valListFile')).read().split()
+ test_list = open(os.path.join(self_dir, extract_dir, 'testListFile')).read().split()
+
+ total_not_found_slot = 0
+ total_slot = 0
+ total_turn = 0
+ total_not_found_turn = 0
+ total_not_found_state = 0
+
+ slot_notfound_dict = {}
+
+ dialog_idx = 0
+ for dialog_id, dialog in tqdm(data.items()):
+
+ acc_not_found_flag = False
+
+ coref_dict = {}
+
+ data_split = None
+ for _split in data_splits:
+ if dialog_id.strip('.json') in eval(_split + '_list'):
+ data_split = _split
+ break
+ # assert data_split is not None
+ # if data_split != 'test':
+ # continue
+ if data_split is None:
+ continue
+
+ if data_split == 'train':
+ num_train_dialogue += len(data)
+
+ dialog_idx += 1
+ # if dialog_idx > 10:
+ # break
+ converted_dialogue = {
+ 'dataset': 'multiwoz21',
+ 'data_split': data_split,
+ 'dialogue_id': 'multiwoz21_' + str(dialog_idx),
+ 'original_id': dialog_id,
+ 'domains': [d for d in dialog['goal'] if
+ len(dialog['goal'][d]) != 0 and d in multiwoz_desc and d not in ['police', 'hospital', 'bus']],
+ 'turns': [],
+ }
+
+ if data_split == 'train':
+ num_train_utt += len(dialog['log'])
+
+ prev_state = None
+ accum_fixed_state = {}
+ for turn_id, turn in enumerate(dialog['log']):
+
+ utt = turn['text'].lower()
+ # for several wrong words
+ utt = utt.replace('seeuni', 'see uni')
+
+ utt = ' '.join(utt.split())
+ utt = utt.replace(' im ', ' i\'m ')
+ utt = utt.replace(' dont ', ' don\'t ')
+ utt = utt.replace(' thats ', ' that\'s ')
+ utt = utt.replace('idon\'t', ' i don\'t ')
+ utt = utt.replace('wedon\'t ', 'we don\'t ')
+ utt = utt.replace('id be ', 'i\'d be ')
+ # utt = utt.replace('cambridgethat\'svery ', 'cambridge that\'s very')
+ utt = re.sub(r'^im ', 'i\'m ', utt)
+ utt = re.sub(r'^whats ', 'what\'s ', utt)
+ utt = re.sub(r'^id ', 'i\'d ', utt)
+ utt = re.sub(r'^thats ', 'that\'s ', utt)
+
+ utt = re.sub( r'([a-zA-Z0-9])([,.!\'-\?"~])', r'\1 \2', utt)
+ utt = re.sub(r'([,.!\'-\?"~])([a-zA-Z0-9])', r'\1 \2', utt)
+
+ das = turn.get('dialog_act', [])
+ role = 'user' if turn_id % 2 == 0 else 'system'
+ spans = turn.get('span_info', [])
+
+ da_dict = {}
+ for Domain_Act in das:
+ Domain = Domain_Act.split('-')[0]
+ if Domain.lower() not in converted_dialogue['domains'] and Domain.lower() not in ['general', 'booking']:
+ continue
+
+ Svs = das[Domain_Act]
+ for S, v in Svs:
+ v = v.lower()
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+
+ if (Domain_Act, S, v) not in da_dict:
+ da_dict[(Domain_Act, S, v)] = {}
+
+ for span in spans:
+ Domain_Act, S, v, start_word, end_word = span
+ v = v.lower()
+ if not (Domain_Act, S, v) in da_dict:
+ # logging.info('span da annotation not found in multiwoz da label')
+ # logging.info(dialog_id, turn_id)
+ # logging.info((Domain_Act, S, v))
+ # logging.info(da_dict)
+ num_match_error_da_span += 1
+ else:
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ da_dict[(Domain_Act, S, v)]['start_word'] = []
+ da_dict[(Domain_Act, S, v)]['end_word'] = []
+
+ da_dict[(Domain_Act, S, v)]['start_word'].append(start_word)
+ da_dict[(Domain_Act, S, v)]['end_word'].append(end_word)
+
+ converted_turn = {
+ 'utt_idx': turn_id,
+ 'speaker': role,
+ 'utterance': utt,
+ 'dialogue_act': convert_da(utt, da_dict, binary_ont, intent_ont, dialog_id, turn_id, da_cat_slot_values),
+ }
+
+ # for state annotations
+ if role == 'system':
+ turn_state = turn['metadata']
+ cur_state = {}
+ for domain in turn_state:
+ if domain in ['police', 'hospital', 'bus']:
+ continue
+ if domain not in converted_dialogue['domains']:
+ continue
+ cur_state[domain] = {}
+ for subdomain in ['semi', 'book']:
+ for slot in turn_state[domain][subdomain]:
+ if slot == 'booked':
+ continue
+ if slot == 'ticket': # or (domain == 'train' and slot == 'people'):
+ # for cases where domain slot exists in REF but not in state
+ # because of check in evaluate.py
+ continue
+
+ else:
+ fixed_slot = slot
+ state_ds = domain + '-' + fixed_slot
+ if state_ds not in slot_to_type:
+ logging.info('state slot not defined in da list')
+ logging.info(state_ds)
+ if turn_state[domain][subdomain][slot] in ['', [], 'not mentioned', 'none']:
+ cur_state[domain][fixed_slot] = ""
+ else:
+ if turn_state[domain][subdomain][slot].startswith('th '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+ if turn_state[domain][subdomain][slot].startswith('he '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+
+ cur_state[domain][fixed_slot] = turn_state[domain][subdomain][slot]
+
+ if domain not in state_ont:
+ state_ont[domain] = []
+ if fixed_slot not in state_ont[domain]:
+ state_ont[domain].append(fixed_slot)
+
+ if domain == 'train' and 'people' not in cur_state[domain]:
+ cur_state[domain]['people'] = ''
+ # if len(converted_turn['state'][domain]) == 0:
+ # converted_turn['state'].pop(domain)
+ if len(converted_dialogue['turns']) > 0:
+ # move state from system side to user side
+ converted_dialogue['turns'][-1]['state'] = copy.deepcopy(cur_state)
+
+ # for state update annotations
+ state_update, _notfoundslot, _totalslot, ret_diff_state = get_state_update(prev_state, cur_state, converted_dialogue,
+ dialog_id, turn_id, turn['text'], coref_dict,
+ slot_notfound_dict, da_cat_slot_values)
+
+ update_state(accum_fixed_state, ret_diff_state)
+ for domain in accum_fixed_state:
+ for slot in accum_fixed_state[domain]:
+ assert isinstance(accum_fixed_state[domain][slot], str), print(accum_fixed_state[domain][slot])
+
+ if _notfoundslot == 0:
+ # for slot in state_update['categorical']:
+ # assert 'fixed_value' not in slot
+ for slot in state_update['non-categorical']:
+ if slot['value'] not in ['', 'dontcare']:
+ assert 'utt_idx' in slot
+
+ else:
+ flag = False
+ for slot in state_update['categorical']:
+ if 'fixed_value' in slot:
+ flag = True
+ break
+ for slot in state_update['non-categorical']:
+ if 'utt_idx' not in slot:
+ flag = True
+ break
+ assert flag, print(flag, state_update['non-categorical'])
+
+ total_turn += 1
+ total_slot += _totalslot
+ total_not_found_slot += _notfoundslot
+ total_not_found_turn += 1 if _notfoundslot > 0 else 0
+ if _notfoundslot > 0:
+ acc_not_found_flag = True
+ if acc_not_found_flag:
+ total_not_found_state += 1
+
+ coref_dict = {}
+ converted_dialogue['turns'][-1]['state_update'] = copy.deepcopy(state_update)
+ converted_dialogue['turns'][-1]['fixed_state'] = copy.deepcopy(accum_fixed_state)
+ if 'state' not in converted_dialogue['turns'][-1]:
+ converted_dialogue['turns'][-1]['state'] = {}
+ prev_state = copy.deepcopy(cur_state)
+
+ converted_dialogue['turns'].append(converted_turn)
+
+ if 'coreference' in turn:
+ for Domain_Act in turn['coreference']:
+ for Slot, value, coref, coref_turn, coref_pos in turn['coreference'][Domain_Act]:
+ value = value.lower()
+ coref_dict[(Domain_Act, Slot, value)] = {'turn': coref_turn, 'pos': coref_pos,
+ 'coref_value': coref,
+ 'utt': converted_dialogue['turns'][coref_turn][
+ 'utterance']}
+
+ check_spans(converted_dialogue)
+ # postprocess_update_spans(converted_dialogue)
+ if converted_dialogue['turns'][-1]['speaker'] == 'system':
+ converted_dialogue['turns'].pop(-1)
+ all_data.append(converted_dialogue)
+
+ print('total_turn', total_turn)
+ print('total_not_found_turn', total_not_found_turn)
+ print('total_slot', total_slot)
+ print('total_not_found_slot', total_not_found_slot)
+ print('total_not_found_state', total_not_found_state)
+ print(slot_notfound_dict)
+ from collections import Counter
+ # print({k : dict(Counter(v)) for k, v in cat_slot_values.items()})
+ json.dump({k : dict(Counter(v)) for k, v in state_cat_slot_values.items()}, open(os.path.join(self_dir, 'cat_slot_values.json'), 'w'), indent=4)
+ cat_slot_values = {k: list(set(v)) for k, v in state_cat_slot_values.items()}
+ da_cat_slot_values = {k: list(set(v)) for k, v in da_cat_slot_values.items()}
+
+ json.dump(all_data, open('data.json', 'w'), indent=4)
+ write_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ os.remove('data.json')
+
+ new_ont = {
+ 'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': {}
+ }
+
+ for d_s in slot_to_type:
+ d, s = d_s.split('-')
+ if d not in new_ont['domains']:
+ new_ont['domains'][d] = {
+ 'description': multiwoz_desc[d]['domain'],
+ 'slots': {}
+ }
+ domain_ont = new_ont['domains'][d]
+ assert s not in domain_ont
+ domain_ont['slots'][s] = {
+ 'description': multiwoz_desc[d][s] if s in multiwoz_desc[d] else '',
+ 'is_categorical': d_s in state_cat_slot_ds,
+ 'possible_values': da_cat_slot_values[d_s] if d_s in state_cat_slot_ds else []
+ }
+ domain_ont['slots'][s]['possible_values'] = [_ for _ in domain_ont['slots'][s]['possible_values'] if _ not in ['dontcare', '']]
+
+ new_ont['state'] = {}
+ # print(state_cat_slot_value_dict)
+ print(state_ont)
+ for d in state_ont:
+ new_ont['state'][d] = {}
+ for s in state_ont[d]:
+ d_s = '-'.join([d, s])
+ new_ont['state'][d][s] = ''
+
+ new_ont['intents'] = {i: {'description': multiwoz_desc['intents'][i]} for i in intent_ont}
+ new_ont['binary_dialogue_act'] = binary_ont
+
+ slot_desc = json.load(open(os.path.join(self_dir, extract_dir, './slot_descriptions.json')))
+ for domain_slot in slot_desc:
+ _domain, _slot = domain_slot.split('-')
+ _desc = slot_desc[domain_slot][0]
+ if _slot == 'arriveby':
+ _slot = 'arriveBy'
+ elif _slot == 'leaveat':
+ _slot = 'leaveAt'
+ if 'book' in _slot:
+ _slot = _slot.replace('book ', '')
+ if not _domain in new_ont['state']:
+ # logging.info('domain {} not in state domains'.format(_domain))
+ continue
+ if _domain in new_ont['domains'] and _slot in new_ont['domains'][_domain]['slots']:
+ new_ont['domains'][_domain]['slots'][_slot]['description'] = _desc
+ if not _slot in new_ont['state'][_domain]:
+ logging.info('domain {} slot {} not in state'.format(_domain, _slot))
+ continue
+ # new_ont['state'][_domain][_slot] = ""
+ assert _domain in new_ont['domains'], print(_domain)
+ assert _slot in new_ont['domains'][_domain]['slots']
+
+ logging.info('num_match_error_da_span {}'.format(num_match_error_da_span))
+ json.dump(new_ont, open(os.path.join(self_dir, './ontology.json'), 'w'), indent=4)
+
+ else:
+ all_data = read_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ new_ont = json.load(open(os.path.join(self_dir, './ontology.json'), 'r'))
+ logging.info('# dialogue: {}, # turn: {}'.format(num_train_dialogue, num_train_utt))
+ return all_data, new_ont
+
+
+# def postprocess_update_spans(dialog):
+# changed_utt_idx_and_position = {}
+# for turn in dialog['turns']:
+# if turn['speaker'] != 'user':
+# continue
+# changed = False
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+#
+# # assume at most one word changes for every utterance
+# if turn['utt_idx'] not in changed_utt_idx_and_position:
+# if utt_idx == turn['utt_idx'] and start-1 > -1 and turn['utterance'][start-1] not in [' ']:
+# changed_utt_idx_and_position[turn['utt_idx']] = start
+# print('=======================')
+# print(dialog['original_id'])
+# print(turn['utterance'])
+# print(json.dumps(_update, indent=2))
+# print(turn['utterance'][start: end])
+# turn['utterance'] = turn['utterance'][:start] + ' ' + turn['utterance'][start:]
+# print(turn['utterance'])
+# _update['start'] += 1
+# _update['end'] += 1
+# changed = True
+# if utt_idx not in changed_utt_idx_and_position:
+# continue
+# else:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update['fixed_value'] != 'not found' else _update['value']
+# if start >= changed_utt_idx_and_position[utt_idx]:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+# _update['start'] += 1
+# _update['end'] += 1
+# elif start < changed_utt_idx_and_position[utt_idx] < end:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1], value)
+# print('fix')
+# print(_update)
+# _update['end'] += 1
+# _update['fixed_value'] = turn['utterance'][_update['start']: _update['end'] + 1].strip()
+# print(_update)
+# if changed:
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+#
+# if utt_idx not in changed_utt_idx_and_position:
+# continue
+# else:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+# 'fixed_value'] != 'not found' else _update['value']
+# if start >= changed_utt_idx_and_position[utt_idx]:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert dialog['turns'][utt_idx]['utterance'][_update['start'] + 1: _update['end'] + 1] == value
+# _update['start'] += 1
+# _update['end'] += 1
+# elif start < changed_utt_idx_and_position[utt_idx] < end:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# print('====================fix===================')
+# print(_update)
+# assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+# _update['end'] += 1
+# _update['fixed_value'] = dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end'] + 1]
+# print(_update)
+# for turn in dialog['turns']:
+# if turn['speaker'] != 'user':
+# continue
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+# 'fixed_value'] != 'not found' else _update['value']
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+# if dialog['turns'][utt_idx]['utterance'][start] == ' ':
+# _update['start'] += 1
+# _update['fixed_value'] = value[1:]
+# value = value[1:]
+# start += 1
+# assert dialog['turns'][utt_idx]['utterance'][start: end] == value, print(json.dumps(turn, indent=4), [c for c in dialog['turns'][utt_idx]['utterance'][start: end]], [c for c in value])
+# return dialog
+
+
+def get_time_variants(time_text):
+ value_list = [time_text]
+ pattern_time = r'(\d{1,2}:\d{2})(\s)?(am|pm|AM|PM)?'
+ match_times = re.findall(pattern_time, time_text)
+ if len(match_times) < 1:
+ return []
+ match_time = match_times[0]
+
+ am_flag = match_time[2] in ['am', 'AM']
+ pm_flag = match_time[2] in ['pm', 'PM']
+ no_am_pm_flag = match_time[2] == ''
+ if am_flag:
+ # 4:00am -> 4:00
+ value_list.append(match_time[0])
+ if len(match_time[0]) == 4:
+ # 4:00 -> 04:00
+ value_list.append('0' + match_time[0])
+ if pm_flag:
+ # 4:00pm -> 16:00
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ new_hour = 12 + hour
+ value_list.append(str(new_hour)+':'+min)
+ if no_am_pm_flag:
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ if hour > 12:
+ new_hour = hour - 12
+ value_list.append(str(new_hour) + ':' + min + 'pm')
+ value_list.append(str(new_hour) + ':' + min + ' pm')
+ value_list.append(str(new_hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(new_hour) + 'pm')
+ value_list.append(str(new_hour) + ' pm')
+ value_list.append(str(new_hour))
+ else:
+ value_list.append(str(hour) + ':' + min + 'am')
+ value_list.append(str(hour) + ':' + min + ' am')
+ value_list.append(str(hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(hour) + 'am')
+ value_list.append(str(hour) + ' am')
+ value_list.append(str(hour))
+ if len(match_time[0]) == 5 and match_time[0][0] == '0':
+ value_list.append(match_time[0][1:])
+ value_list.append(''.join(match_time[0].split(':')))
+
+ return value_list
+
+
+def get_genitive_variants(value):
+ ret_list = []
+ value_genitive_format = r"(?=\w)s(?=\s)"
+ value_pattern = re.compile(value_genitive_format)
+
+ span_genitive_value = re.sub(value_pattern, " 's", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ span_genitive_value = re.sub(value_pattern, "'s", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ # if len(ret_list) > 0:
+ # print('=============================')
+ # print(value)
+ # print(re.findall(value_pattern, value))
+ # print(ret_list)
+ return ret_list
+
+
+def check_spans(dialog):
+ for turn in dialog['turns']:
+ if turn['speaker'] != 'user':
+ continue
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+ 'fixed_value'] != 'not found' else _update['value']
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+ assert dialog['turns'][utt_idx]['utterance'][start:end] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][start:end])
+
+
+
+def get_bb_variants(value):
+ ret_list = []
+ if 'bed and breakfast' in value:
+ ret_list.append(value.replace('bed and breakfast', 'b & b'))
+ return ret_list
+
+if __name__ == '__main__':
+ preprocess(da_cat_slot_values, state_cat_slot_values)
\ No newline at end of file
diff --git a/data/unified_datasets/multiwoz22/README.md b/data/unified_datasets/multiwoz22/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..52db0dd3d3e5e16946d770bc3484d17b50b0dc5c
--- /dev/null
+++ b/data/unified_datasets/multiwoz22/README.md
@@ -0,0 +1,34 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots. state and state updates.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 8434 | 105066 | 12.46 | 17.27 | 7 |
+| dev | 999 | 13731 | 13.74 | 17.72 | 7 |
+| train | 1000 | 13744 | 13.74 | 17.67 | 7 |
+
+
+## Main changes
+
+- only keep 5 domains in state annotations and dialog acts.
+- `pricerange`, `area`, `day`, `internet`, `parking`, `stars` are considered categorical slots.
+- punctuation marks are split from their previous tokens. e.g `I want to find a hotel. ->
+ I want to find a hotel .`
+
+Run `evaluate.py`:
+
+da values match rate: 97.944
+state values match rate: 66.945
+
+### original data
+
+- from [multiwoz](https://github.com/budzianowski/multiwoz) repo.
+- original multiwoz2.2 dataset gives slot value in List format. We take the first value
+in each slot list as ground-truth value.
+
+
diff --git a/data/unified_datasets/multiwoz22/data.zip b/data/unified_datasets/multiwoz22/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..92d82a3382228455b61a4117413731d17ee7add2
Binary files /dev/null and b/data/unified_datasets/multiwoz22/data.zip differ
diff --git a/data/unified_datasets/multiwoz22/ontology.json b/data/unified_datasets/multiwoz22/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..d3cc3c187cc961994cfb477f61a11694c6f156fd
--- /dev/null
+++ b/data/unified_datasets/multiwoz22/ontology.json
@@ -0,0 +1,1879 @@
+{
+ "domains": {
+ "taxi": {
+ "description": "taxi information query system",
+ "slots": {
+ "destination": {
+ "description": "destination of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_phone": {
+ "description": "taxi phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_types": {
+ "description": "taxi type",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "restaurant": {
+ "description": "restaurant information query system",
+ "slots": {
+ "food": {
+ "description": "food type for the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget for the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "varying price",
+ "every price point",
+ "a little expensive but worth it",
+ "assorted",
+ "most affordable price range",
+ "the least expensive",
+ "same price range",
+ "similar price range",
+ "priced cheaply",
+ "varying price range",
+ "any price point",
+ "cheap priced",
+ "moderate price range",
+ "moderate",
+ "any",
+ "rather expensive",
+ "that price range",
+ "on the pricey side",
+ "on the cheap side",
+ "various",
+ "your price range",
+ "expensively priced",
+ "expensive or moderate",
+ "moderatley priced",
+ "moderatre",
+ "more expensive",
+ "adforable",
+ "fine",
+ "not moderate",
+ "moderately expensive",
+ "relatively cheap",
+ "cheap or expensive",
+ "moderately prices",
+ "expensive side",
+ "ranging from cheap to expensive",
+ "very expensive",
+ "ch",
+ "cherap",
+ "do nt care",
+ "low cost",
+ "do n't care",
+ "most expensive",
+ "or the moderately priced",
+ "moderate and expensive",
+ "that price",
+ "expensively",
+ "reasonable",
+ "not cheap",
+ "moderately priced",
+ "all price ranges",
+ "or otherwise",
+ "upscale",
+ "expensive / upscale",
+ "fairly expensive",
+ "great prices",
+ "quite low",
+ "moderatly",
+ "moderate or expensive priced",
+ "cheaply priced",
+ "the same",
+ "expensive range",
+ "moderate priced",
+ "this price range",
+ "pricey",
+ "moderately - priced",
+ "reasonably priced",
+ "high - end",
+ "cheaper",
+ "expensive and moderately priced",
+ "pretty expensive",
+ "modest",
+ "they vary",
+ "luxury options",
+ "fairly inexpensive",
+ "pretty cheap",
+ "expensive but worth every penny",
+ "moderate and one in the cheap range",
+ "not expensive",
+ "that range",
+ "low priced",
+ "relatively expensive",
+ "cheap to expensive",
+ "mostly expensive and moderately priced",
+ "not - so - expensive",
+ "quite inexpensive",
+ "inexpensive",
+ "your",
+ "budget - friendly",
+ "high class",
+ "expinsive",
+ "fairly cheap",
+ "expensive price range",
+ "any price range",
+ "moderate pricing",
+ "high end",
+ "epensive",
+ "vietnamese",
+ "cheap price range",
+ "mostly expensive",
+ "less expensive",
+ "moderate range",
+ "more moderate",
+ "priced moderately",
+ "expensive price",
+ "centre",
+ "expensive",
+ "middle",
+ "boderate",
+ "moderate price",
+ "moderately",
+ "moderatly priced",
+ "cheap range",
+ "regardless of price",
+ "moderately price",
+ "north",
+ "not too expensive",
+ "not too pricey",
+ "affordable",
+ "all",
+ "east",
+ "quite expensive",
+ "cheaply - priced",
+ "budget conscious",
+ "cheap",
+ "budget friendly"
+ ]
+ },
+ "address": {
+ "description": "exact location of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "center of town",
+ "west part of town .",
+ "closely located",
+ "south cambridge",
+ "cetre",
+ "the city centre",
+ "on the east",
+ "any part of town",
+ "centre area of the city",
+ "the north",
+ "near centre",
+ "center",
+ "central region of town",
+ "not",
+ "east area of town",
+ "north side of cambridge",
+ "north of town",
+ "centre area",
+ "north part of town",
+ "west part of town",
+ "the south area of the city",
+ "here",
+ "south part",
+ "west",
+ "in town",
+ "in the centre",
+ "the southern area",
+ "south side",
+ "west area",
+ "anywhere",
+ "west cambridge",
+ "north area of town",
+ "the east side of town",
+ "any",
+ "south area of town",
+ "centre cambridge",
+ "centre of town",
+ "this town",
+ "the east part of town",
+ "cambridge",
+ "northern part",
+ "eastside",
+ "centrally",
+ "around town",
+ "close to your location",
+ "westies",
+ "the north area",
+ "thai",
+ "centre of town .",
+ "in town at all",
+ "in the west",
+ "mexican",
+ "not in the west",
+ "the town centre",
+ "other part of the town",
+ "the centre",
+ "southern area",
+ "the west",
+ "any area",
+ "southend",
+ "south of cambridge",
+ "east section of town",
+ "the north side",
+ "the east",
+ "the north side .",
+ "elsewhere in the city",
+ "south of town .",
+ "the west side of town",
+ "the centre area of town",
+ "the center part of town",
+ "northside",
+ "do nt care",
+ "that area",
+ "the are",
+ "do n't care",
+ "north end",
+ "in the city",
+ "centrem",
+ "the centre area",
+ "south of town",
+ "the east side",
+ "centre city",
+ "the south of town",
+ "the north end",
+ "centre region",
+ "the south side of town",
+ "east part of tow",
+ "the center",
+ "the same area",
+ "east side",
+ "town",
+ "any where in the city",
+ "throughout the area",
+ "in the city .",
+ "town centre",
+ "same side of town as your hotel",
+ "the south area of town",
+ "and centre",
+ "east area",
+ "north area",
+ "the west end",
+ "north cambridge",
+ "the area you have chosen",
+ "west side",
+ "east of town",
+ "central",
+ "east part of town",
+ "that part of town",
+ "city center",
+ "the same",
+ "the south side",
+ "city",
+ "by the airport",
+ "downtown",
+ "the center of town",
+ "chesterton",
+ "east section",
+ "southern",
+ "south part of time",
+ "ely",
+ "within this area",
+ "the north part of town",
+ "north or centre",
+ "the west side",
+ "all over the city",
+ "center of the town",
+ "cambridge centre",
+ "wet part of town",
+ "center area of town",
+ "west end of town",
+ "city centre",
+ "the area",
+ "all of cambridge",
+ "towns centre",
+ "in the north",
+ "west part of the city",
+ "north side of town",
+ "the south part of town",
+ "centre of the city",
+ "the east area",
+ "here in the city",
+ "that area of town",
+ "east side of the city",
+ "centre of cambridge",
+ "the city center",
+ "almost every area of town",
+ "around there",
+ "north side of chersteron",
+ "in cambridge",
+ "west area of town",
+ "the area of west",
+ "town centre area",
+ "nearby",
+ "same area as the hotel",
+ "south part of town",
+ "not in the south of town",
+ "centreof",
+ "ctre",
+ "anywhere in cambridge",
+ "this area",
+ "town of centre",
+ "town center",
+ "in the east",
+ "east of cambridge",
+ "southside",
+ "all over town",
+ "south side of town",
+ "other parts of town",
+ "northern part of cambridge",
+ "the same area as the botanic gardens",
+ "center cambridge",
+ "south area",
+ "south",
+ "near the centre",
+ "east end",
+ "throughout the city",
+ "the centre of town",
+ "centere",
+ "near clare hall",
+ "close to that area",
+ "in the city centre",
+ "n the centre",
+ "the centre part of town",
+ "the west of town",
+ "around the college",
+ "centre area of town",
+ "centre",
+ "these areas",
+ "south end",
+ "close to the hotel",
+ "expensive",
+ "south part of the city",
+ "close it city centre",
+ "west of cambridge",
+ "same area as the park",
+ "east side of town",
+ "that",
+ "west end",
+ "the south",
+ "all over cambridge",
+ "north",
+ "anywhere in town",
+ "centre area of the town",
+ "the west part of town",
+ "same area",
+ "south side of the town",
+ "the west area",
+ "near the center of town",
+ "centre part of town",
+ "northern parts of town",
+ "west side of town",
+ "east",
+ "westside",
+ "west of town",
+ "the north side of town",
+ "north and west",
+ "centrally located",
+ "west side area",
+ "same",
+ "central area",
+ "cheap",
+ "north side"
+ ]
+ },
+ "postcode": {
+ "description": "postcode of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "restaurant phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "time of the restaurant booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the restaurant booking",
+ "is_categorical": true,
+ "possible_values": [
+ "w",
+ "thursday",
+ "monday",
+ "friday",
+ "saturday",
+ "tuesday",
+ "wednesday",
+ "sunday"
+ ]
+ },
+ "choice": {
+ "description": "number of restaurants meeting requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "attraction": {
+ "description": "an entertainment that is offered to the public",
+ "slots": {
+ "address": {
+ "description": "details of where the attraction is",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "phone number of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "entrance fee": {
+ "description": "the fee charged for admission to the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "type of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the attraction",
+ "is_categorical": true,
+ "possible_values": [
+ "center of town",
+ "that side of town",
+ "south cambridge",
+ "the city centre",
+ "north in milton",
+ "western part of the town",
+ "western area of town",
+ "whole of cambridge",
+ "the south near your hotel",
+ "the north",
+ "your desired location",
+ "center",
+ "cenre",
+ "east area of town",
+ "north part of town",
+ "centre area",
+ "north of town",
+ "west part of town",
+ "near ely",
+ "center part of town",
+ "western",
+ "west",
+ "in town",
+ "in the centre",
+ "south side",
+ "west cambridge",
+ "northern area",
+ "west area",
+ "the north of the city",
+ "centre area of cambridge",
+ "western cambridge",
+ "western part of town",
+ "centre of town",
+ "the east part of town",
+ "cambridge",
+ "various",
+ "west area of the city",
+ "eastside",
+ "eat",
+ "in the south",
+ "cetnre of town",
+ "right in the center of town",
+ "northern",
+ "different area",
+ "another area",
+ "centre area .",
+ "centre by galleria",
+ "same general area",
+ "the centre",
+ "clifton way",
+ "the west",
+ "east cambridge",
+ "close to the restaurant",
+ "this side of town",
+ "the east of the town",
+ "south of cambridge",
+ "most of them are in the center",
+ "the east",
+ "same area as tandoori palace",
+ "went side of town",
+ "the west side of town",
+ "do nt care",
+ "close to the center of town",
+ "that area",
+ "central zone",
+ "the centre area",
+ "south of town",
+ "near the restaurant",
+ "the east side",
+ "centre city",
+ "centre region",
+ "center of town .",
+ "center area",
+ "all of the other areas",
+ "the same area",
+ "the center",
+ "east side",
+ "central cambridge",
+ "town centre",
+ "center are",
+ "south park of town",
+ "center of cambridge",
+ "east area",
+ "northern area of town",
+ "north area",
+ "the west end",
+ "east of town",
+ "to the south",
+ "west side",
+ "central",
+ "that part of town",
+ "east part of town",
+ "city center",
+ "west side of the city",
+ "cambridge 's centre",
+ "the south side",
+ "city cenre",
+ "city",
+ "downtown",
+ "in that area",
+ "cb30aq",
+ "east section",
+ "ely",
+ "the area you 're looking for",
+ "the west side",
+ "all over the city",
+ "center of the town",
+ "the eastside",
+ "in the area",
+ "other areas",
+ "the west area of town",
+ "in the center",
+ "west end of town",
+ "city centre",
+ "city centre .",
+ "entre",
+ "the area",
+ "towns centre",
+ "northend",
+ "west part of the city",
+ "the center area",
+ "near the hotel",
+ "centre of town ?",
+ "south part of cambridge",
+ "that area .",
+ "centre of the city",
+ "southern cambridge",
+ "museum",
+ "centre of cambridge",
+ "all around the city",
+ "cambridge leisure park",
+ "several different parts of town",
+ "city 's centre",
+ "center of the city",
+ "in cambridge",
+ "church area",
+ "museums",
+ "west end of the city",
+ "west area of town",
+ "same part",
+ "town centre area",
+ "nearby",
+ "south part of town",
+ "same part of town as your restaurant",
+ "the same road",
+ "this area",
+ "in the east",
+ "town center",
+ "there",
+ "same area as hotel",
+ "ce",
+ "east of cambridge",
+ "south side of town",
+ "other parts of town",
+ "all over town",
+ "central district",
+ "the west - side",
+ "the center of the park",
+ "south area",
+ "south",
+ "near the centre",
+ "throughout the city",
+ "the centre of town",
+ "the south area",
+ "in that side",
+ "very close in the same area",
+ "cent",
+ "wet end",
+ "the centre part of town",
+ "west of your city",
+ "the west of town",
+ "we",
+ "north section of cambridge",
+ "centre area of town",
+ "centre",
+ "norwich",
+ "south end",
+ "every area except the north",
+ "the centry area",
+ "east side of town",
+ "that",
+ "northern cambridge",
+ "west end",
+ "the south",
+ "north of the city",
+ "north",
+ "east end of town",
+ "the center part of town",
+ "on the centre",
+ "same area",
+ "the west area",
+ "centre part of town",
+ "west side of town",
+ "all",
+ "east",
+ "westside",
+ "centrally located",
+ "same",
+ "the area you are looking for",
+ "west part of cambridge",
+ "north side"
+ ]
+ },
+ "name": {
+ "description": "name of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "the price range for the attraction, from cheap to expensive",
+ "is_categorical": true,
+ "possible_values": [
+ "neither prices are listed",
+ "from free to 2 pounds 50",
+ "entrance fee",
+ "5 pounds",
+ "4 pounds",
+ "unaware of what their entrance fee is",
+ "do n't see a price listed",
+ "expensive",
+ "they do n't have the entrance fee posted",
+ "no entrance fee",
+ "free",
+ "5 pound entrance fee",
+ "2 pounds",
+ "not sure of the fee",
+ "we do n't have any information",
+ "free admission",
+ "do n't have information",
+ "3.50 pounds",
+ "moderately priced",
+ "cheap",
+ "not listed"
+ ]
+ },
+ "choice": {
+ "description": "number of attractions matching requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "booking": {
+ "description": "to arrange with a taxi, restaurant, train, etc.",
+ "slots": {
+ "time": {
+ "description": "time for an order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day for an order, from monday to sunday",
+ "is_categorical": true,
+ "possible_values": [
+ "tues",
+ "monday",
+ "saturday night thru monday",
+ "tuesday",
+ "tuesday through friday",
+ "tuesday instead of thursday",
+ "tuesday 's",
+ "sunday",
+ "sundar",
+ "sunday 18/06/2017",
+ "tonight",
+ "this evening",
+ "sunday and monday",
+ "vmhj6y3i",
+ "at that time",
+ "today",
+ "friday",
+ "saturday",
+ "that",
+ "thursday",
+ "we d",
+ "weds",
+ "2",
+ "the same day",
+ "6",
+ "1",
+ "wednesday",
+ "that time frame"
+ ]
+ },
+ "stay": {
+ "description": "for how long the user wish to be at a place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "how many person the order is for",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the ordered place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "train": {
+ "description": "query and order a train",
+ "slots": {
+ "destination": {
+ "description": "destination of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "duration": {
+ "description": "the length of time the train trip lasts",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking for train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the train",
+ "is_categorical": true,
+ "possible_values": [
+ "monday",
+ "tuesday",
+ "tr2519",
+ "sunday",
+ "between friday and wednesday",
+ "sundays",
+ "every",
+ "daily",
+ "mondays",
+ "sat",
+ "we",
+ "all week",
+ "every day",
+ "friday",
+ "saturday",
+ "that",
+ "train",
+ "other days are available",
+ "this day",
+ "frday",
+ "cambridge",
+ "thursday",
+ "weds",
+ "saturdays",
+ "that day",
+ "13:07",
+ "fiday",
+ "everday",
+ "wednesday",
+ "fr"
+ ]
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time for the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "price for the train ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "choice": {
+ "description": "number of trains that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel": {
+ "description": "to query hotel information and place an order",
+ "slots": {
+ "internet": {
+ "description": "internet option at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "yes",
+ "4",
+ "do nt care",
+ "no",
+ "free",
+ "y",
+ "free internet",
+ "none"
+ ]
+ },
+ "area": {
+ "description": "area or place of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "center of town",
+ "that side of town",
+ "south cambridge",
+ "several areas of town",
+ "any part of the city",
+ "any part of town",
+ "the east end of town",
+ "a different area",
+ "the north",
+ "different part of town",
+ "north end of the city",
+ "center",
+ "and east",
+ "near centre",
+ "norht",
+ "northern side of town",
+ "east area of town",
+ "somewhere else",
+ "north of town",
+ "centre area",
+ "north part of town",
+ "west part of town",
+ "east and the north .",
+ "next door",
+ "all over",
+ "west",
+ "in town",
+ "in the centre",
+ "the town center",
+ "west part",
+ "south side",
+ "west area",
+ "northern area",
+ "the north of the city",
+ "north area of town",
+ "western cambridge",
+ "the east side of town",
+ "different parts of town",
+ "south area of town",
+ "centre cambridge",
+ "centre of town",
+ "cambridge",
+ "various",
+ "on the west side",
+ "north end of town",
+ "eastside",
+ "centrally",
+ "near the museum",
+ "eat",
+ "el shaddai",
+ "on the south",
+ "close to where you 'll be dining",
+ "in the west",
+ "northern",
+ "the town centre",
+ "different area",
+ "the northern part of town",
+ "the centre",
+ "the west",
+ "east cambridge",
+ "any area",
+ "southend",
+ "east section of town",
+ "the north side",
+ "southern part of town",
+ "the east",
+ "south end of town",
+ "west side near the restaurant",
+ "northside",
+ "do nt care",
+ "that area",
+ "north end",
+ "in the city",
+ "the centre area",
+ "south of town",
+ "near the restaurant",
+ "the east side",
+ "the city 's south side",
+ "centre city",
+ "the north end",
+ "north park of town",
+ "all areas",
+ "all parts of the city",
+ "south cambridge area",
+ "other parts of the city",
+ "that region",
+ "the same area",
+ "centra",
+ "east side",
+ "town centre",
+ "and in the centre",
+ "same area as the restaurant",
+ "north part of the city",
+ "across cambridge",
+ "east area",
+ "north area",
+ "the west end",
+ "north cambridge",
+ "east of town",
+ "west side",
+ "everywhere but the city centre",
+ "east part of town",
+ "that part of town",
+ "except in the north",
+ "city center",
+ "across town",
+ "west areas of town",
+ "the south side",
+ "the same",
+ "north part of town .",
+ "north location",
+ "different parts of the city",
+ "on the west",
+ "north of cambridge",
+ "downtown",
+ "the center of town",
+ "the westside",
+ "the north part of town",
+ "all over the city",
+ "the west side",
+ "center of the town",
+ "that vicinity",
+ "on the north",
+ "in the area",
+ "in the center",
+ "city centre",
+ "the easy",
+ "northe part of town",
+ "the area",
+ "all of cambridge",
+ "in the north",
+ "northend",
+ "or west ?",
+ "west part of the city",
+ "north side of town",
+ "same side of town",
+ "centre of the city",
+ "the east area",
+ "except in the east",
+ "place to be a guesthouse",
+ "that area of town",
+ "centre of cambridge",
+ "all around the city",
+ "in cambridge",
+ "the centre of cambridge",
+ "west end of the city",
+ "centre part",
+ "north par of town",
+ "central location",
+ "on the eastside",
+ "south part of town",
+ "eastern part of the city",
+ "the centre of the city",
+ "east part of time",
+ "anywhere in cambridge",
+ "there",
+ "north and centre",
+ "south side of town",
+ "all over town",
+ "other parts of town",
+ "southside",
+ "nborth",
+ "you are interested in",
+ "south area",
+ "south",
+ "near the centre",
+ "throughout the city",
+ "the centre of town",
+ "the south area",
+ "south closer to the museum",
+ "we",
+ "not in the south",
+ "centre area of town",
+ "centre",
+ "wast",
+ "south end",
+ "eastern cambridge",
+ "various parts of the city",
+ "east side of town",
+ "northern cambridge",
+ "eastern",
+ "west end",
+ "the south",
+ "north",
+ "east near other shops and boutiques",
+ "same area",
+ "the west area",
+ "centre part of town",
+ "all across town",
+ "west side of town",
+ "throughout cambridge",
+ "another area of town",
+ "south end of the city",
+ "east",
+ "northern part of town",
+ "nearby the restaurant",
+ "west of town",
+ "centrally located",
+ "centrally - located",
+ "cheap",
+ "not to far from the restaurant",
+ "north side"
+ ]
+ },
+ "stars": {
+ "description": "star rating of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "3-star",
+ "3",
+ "does not have",
+ "foru",
+ "1-star",
+ "not as fancy",
+ "four star",
+ "yes",
+ "four starts",
+ "do nt care",
+ "five",
+ "one",
+ "no star rating",
+ "two",
+ "5-star",
+ "several",
+ "0 to 4",
+ "four - star",
+ "four stars",
+ "no",
+ "0",
+ "0-star",
+ "2-star",
+ "5",
+ "drop the star rating",
+ "different star ratings",
+ "ranging from 2 - 4 stars",
+ "2",
+ "lower",
+ "4-star",
+ "4",
+ "three",
+ "zero",
+ "1",
+ "does not show",
+ "unrated",
+ "not rated",
+ "four"
+ ]
+ },
+ "parking": {
+ "description": "parking facility at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "yes",
+ "free parking",
+ "do nt care",
+ "no",
+ "free",
+ "n",
+ "none"
+ ]
+ },
+ "phone": {
+ "description": "hotel phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "different",
+ "different price range",
+ "cheapest",
+ "budget - priced",
+ "cheaper than the others",
+ "same price range",
+ "cheap priced",
+ "moderate price range",
+ "extremely reasonable",
+ "slightly more expensive",
+ "other ranges",
+ "same price",
+ "moderate",
+ "any",
+ "moderatly price",
+ "rather expensive",
+ "that price range",
+ "on the cheap side",
+ "your price range",
+ "lower",
+ "expensively priced",
+ "does not say",
+ "moderatley priced",
+ "cheaply",
+ "more expensive",
+ "expensive to moderate",
+ "moderately pried",
+ "great",
+ "relatively cheap",
+ "moderately prices",
+ "moderately priceed",
+ "unfortunately do not have the price",
+ "very cheap",
+ "do nt care",
+ "moderately priced .",
+ "low cost",
+ "all different price ranges",
+ "varying price ranges",
+ "or expensive ?",
+ "very inexpensive",
+ "moderate to cheap",
+ "even cheaper",
+ "moderate in price",
+ "less costly",
+ "super cheap",
+ "moderately priced",
+ "cheap side",
+ "higher price range",
+ "upscale",
+ "espensive",
+ "more moderately priced",
+ "cheaper side",
+ "cheaply priced",
+ "the same",
+ "expensive range",
+ "moderate priced",
+ "your chosen",
+ "economical",
+ "moderately - priced",
+ "different price ranges",
+ "modrate",
+ "your price",
+ "more budget - friendly",
+ "quite cheap",
+ "mostly in the cheap to moderate price range",
+ "cheaper",
+ "economically priced",
+ "lower end",
+ "affordable pricing",
+ "pretty expensive",
+ "cheap to moderate",
+ "oderately priced",
+ "epxensive",
+ "cheap or moderate",
+ "pretty cheap",
+ "very affordable",
+ "free",
+ "not expensive",
+ "expensive or cheap",
+ "that range",
+ "much cheaper",
+ "cheap to expensive",
+ "2",
+ "unknown",
+ "inexpensive",
+ "fairly cheap",
+ "ca n't view that information",
+ "expensive price range",
+ "a little pricey",
+ "chear",
+ "moderately to expensively priced",
+ "moderate pricing",
+ "moderately pricing",
+ "moderate prices",
+ "on the more expensive side",
+ "cheap price range",
+ "the cheapest",
+ "moderate range",
+ "priced moderately",
+ "centre",
+ "expensive",
+ "inexpensively - priced",
+ "guesthouses",
+ "moderate price",
+ "moderately",
+ "moderate price point",
+ "that",
+ "moderatly priced",
+ "expensively - priced",
+ "cheap range",
+ "moderate to cheap range",
+ "moderately price",
+ "north",
+ "affordable",
+ "moderate or cheap",
+ "fairly cheap compared to other hotels",
+ "all",
+ "quite expensive",
+ "cheaply - priced",
+ "cheaply prices",
+ "moderately - priced or cheap",
+ "cheap",
+ "hotel",
+ "hotels"
+ ]
+ },
+ "people": {
+ "description": "number of people for the hotel booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "exact location of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "what is the type of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "stay": {
+ "description": "length of stay at the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the hotel booking",
+ "is_categorical": true,
+ "possible_values": [
+ "thursday",
+ "monday",
+ "t",
+ "friday",
+ "next friday",
+ "saturday",
+ "tuesday",
+ "monda",
+ "wednesday",
+ "sunday"
+ ]
+ },
+ "choice": {
+ "description": "number of hotels that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "police": {
+ "description": "find police stations",
+ "slots": {
+ "address": {
+ "description": "exact location of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "police station phone number",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "inform user of value for a certain slot"
+ },
+ "request": {
+ "description": "ask for value of a slot"
+ },
+ "recommend": {
+ "description": "recommend a choice for user request"
+ },
+ "book": {
+ "description": "place an order for user"
+ },
+ "nobook": {
+ "description": "inform user of booking failure"
+ },
+ "thank": {
+ "description": "express gratitude"
+ },
+ "welcome": {
+ "description": "welcome"
+ },
+ "bye": {
+ "description": "end a conversation and say goodbye to user"
+ },
+ "reqmore": {
+ "description": "ask user for more instructions"
+ },
+ "select": {
+ "description": "provide several choices for user to choose from"
+ },
+ "nooffer": {
+ "description": "inform user that no options matches user request"
+ },
+ "greet": {
+ "description": "express greeting"
+ },
+ "offerbook": {
+ "description": "offer to place an order for user"
+ },
+ "offerbooked": {
+ "description": "inform user that an order is succussful"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "stay",
+ "value": ""
+ },
+ {
+ "intent": "nobook",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "thank",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "welcome",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "food",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "stars",
+ "value": ""
+ },
+ {
+ "intent": "bye",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "reqmore",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "parking",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "taxi",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_types",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "book",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "greet",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "internet",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "offerbook",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "Ref",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "offerbooked",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "entrance fee",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "state": {
+ "hotel": {
+ "area": "",
+ "internet": "",
+ "name": "",
+ "parking": "",
+ "pricerange": "",
+ "stars": "",
+ "type": "",
+ "day": "",
+ "people": "",
+ "stay": ""
+ },
+ "restaurant": {
+ "area": "",
+ "food": "",
+ "name": "",
+ "pricerange": "",
+ "day": "",
+ "people": "",
+ "time": ""
+ },
+ "taxi": {
+ "arriveBy": "",
+ "departure": "",
+ "destination": "",
+ "leaveAt": ""
+ },
+ "train": {
+ "arriveBy": "",
+ "day": "",
+ "departure": "",
+ "destination": "",
+ "leaveAt": "",
+ "people": ""
+ },
+ "attraction": {
+ "area": "",
+ "name": "",
+ "type": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/multiwoz22/original_data.zip b/data/unified_datasets/multiwoz22/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..6fdaeb71ece04f47b6d1b0aa61e1b1ef6a1940fc
Binary files /dev/null and b/data/unified_datasets/multiwoz22/original_data.zip differ
diff --git a/data/unified_datasets/multiwoz22/preprocess.py b/data/unified_datasets/multiwoz22/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..521facf953fd4a4ed4222e4ce9084a0a90a1b518
--- /dev/null
+++ b/data/unified_datasets/multiwoz22/preprocess.py
@@ -0,0 +1,1431 @@
+import copy
+import re
+import zipfile
+import json
+import os
+from tqdm import tqdm
+import sys
+import difflib
+from fuzzywuzzy import fuzz
+sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
+
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+import logging
+
+
+logging.basicConfig(level=logging.INFO)
+self_dir = (os.path.abspath(os.getcwd()))
+
+REF_SYS_DA = {
+ 'Attraction': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Fee': "entrance fee", 'Name': "name", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange", 'Type': "type",
+ 'none': None, 'Open': None
+ },
+ 'Hospital': {
+ 'Department': 'department', 'Addr': 'address', 'Post': 'postcode',
+ 'Phone': 'phone', 'none': None
+ },
+ 'Booking': {
+ 'Day': 'day', 'Name': 'name', 'People': 'people',
+ 'Ref': 'Ref', 'Stay': 'stay', 'Time': 'time',
+ 'none': None
+ },
+ 'Hotel': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Internet': "internet", 'Name': "name", 'Parking': "parking",
+ 'Phone': "phone", 'Post': "postcode", 'Price': "pricerange",
+ 'Stars': "stars", 'Type': "type", 'Stay': 'stay', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Restaurant': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Name': "name", 'Food': "food", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange",
+ 'Time': 'time', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Taxi': {
+ 'Arrive': "arriveBy", 'Car': "taxi_types", 'Depart': "departure",
+ 'Dest': "destination", 'Leave': "leaveAt", 'Phone': "taxi_phone",
+ 'none': None
+ },
+ 'Train': {
+ 'Arrive': "arriveBy", 'Choice': "choice", 'Day': "day",
+ 'Depart': "departure", 'Dest': "destination",
+ 'Leave': "leaveAt", 'People': "people", 'Ref': "Ref",
+ 'Time': "duration", 'none': None, 'Ticket': 'price',
+ },
+ 'Police': {
+ 'Addr': "address", 'Post': "postcode", 'Phone': "phone", 'none': None
+ },
+}
+
+# taxi restaurant attraction train
+slot_to_type = {
+ 'taxi-destination': 'non',
+ 'taxi-departure': 'non',
+ 'taxi-leaveAt': 'non',
+ 'taxi-arriveBy': 'non',
+ 'restaurant-food': 'non',
+ 'restaurant-name': 'non',
+ 'attraction-address': 'non',
+ 'attraction-postcode': 'non',
+ 'restaurant-pricerange': 'cat',
+ 'restaurant-address': 'non',
+ 'restaurant-area': 'cat',
+ 'restaurant-postcode': 'non',
+ 'attraction-phone': 'non',
+ 'attraction-entrance fee': 'non',
+ 'booking-time': 'non',
+ 'booking-day': 'cat',
+ 'attraction-type': 'non',
+ 'attraction-area': 'cat',
+ 'train-destination': 'non',
+ 'train-arriveBy': 'non',
+ 'train-departure': 'non',
+ 'hotel-internet': 'cat',
+ 'hotel-area': 'cat',
+ 'booking-stay': 'non',
+ 'booking-people': 'non',
+ 'train-duration': 'non',
+ 'train-people': 'non',
+ 'train-day': 'cat',
+ 'train-Ref': 'non',
+ 'hotel-stars': 'cat',
+ 'train-leaveAt': 'non',
+ 'train-price': 'non',
+ 'hotel-parking': 'cat',
+ 'hotel-phone': 'non',
+ 'hotel-name': 'non',
+ 'hotel-pricerange': 'cat',
+ 'hotel-people': 'non',
+ 'restaurant-phone': 'non',
+ 'hotel-postcode': 'non',
+ 'hotel-address': 'non',
+ 'attraction-name': 'non',
+ 'hotel-type': 'non',
+ 'restaurant-people': 'non',
+ 'train-choice': 'non',
+ 'attraction-pricerange': 'cat',
+ 'hotel-stay': 'non',
+ 'booking-name': 'non',
+ 'booking-Ref': 'non',
+ 'restaurant-time': 'non',
+ 'restaurant-day': 'cat',
+ 'hotel-day': 'cat',
+ 'hotel-choice': 'non',
+ 'restaurant-choice': 'non',
+ 'attraction-choice': 'non',
+ 'taxi-taxi_phone': 'non',
+ 'taxi-taxi_types': 'non',
+ 'police-address': 'non',
+ 'police-postcode': 'non',
+ 'police-phone': 'non'
+}
+
+state_cat_slot_value_dict = {
+ "hotel-pricerange": {
+ "cheap": 735,
+ "moderate": 1063,
+ "expensive": 594,
+ },
+ "hotel-parking": {
+ "yes": 1809,
+ "no": 126,
+ "free": 4,
+ },
+ "hotel-day": {
+ "tuesday": 385,
+ "wednesday": 410,
+ "monday": 365,
+ "saturday": 407,
+ "friday": 393,
+ "thursday": 384,
+ "sunday": 369,
+ },
+ "train-day": {
+ "wednesday": 533,
+ "monday": 533,
+ "saturday": 543,
+ "thursday": 547,
+ "friday": 563,
+ "tuesday": 553,
+ "sunday": 613,
+ },
+ "hotel-stars": {
+ "4": 1263,
+ "2": 193,
+ "0": 201,
+ "3": 401,
+ "5": 45,
+ "1": 45,
+ },
+ "hotel-internet": {
+ "yes": 1841,
+ "no": 79,
+ "free": 2
+ },
+ "hotel-area": {
+ "east": 416,
+ "north": 717,
+ "centre": 538,
+ "south": 289,
+ "west": 316,
+ },
+ "attraction-area": {
+ "centre": 1290,
+ "west": 332,
+ "north": 155,
+ "south": 240,
+ "east": 272,
+ },
+ "restaurant-pricerange": {
+ "expensive": 1477,
+ "cheap": 758,
+ "moderate": 1028,
+ },
+ "restaurant-area": {
+ "centre": 1745,
+ "south": 398,
+ "north": 390,
+ "east": 360,
+ "west": 423,
+ },
+ "restaurant-day": {
+ "thursday": 362,
+ "wednesday": 412,
+ "friday": 395,
+ "monday": 383,
+ "sunday": 399,
+ "saturday": 421,
+ "tuesday": 350,
+ }
+}
+
+
+synonyms = [
+ ["el shaddia guesthouse", "el shaddai"],
+ [ "peterborough", "peterbourgh"],
+ ["night club", "nightclub", 'nightclubs'],
+ ["boat", "boating"],
+ ["portugese", "portuguese"],
+ ["guesthouse", "guest house"],
+ ["seafood", "sea food"],
+ ["christ 's college", "christ college"],
+ ["huntingdon marriott hotel"]
+]
+
+state_cat_slot_ds = [k for k, v in slot_to_type.items() if v == 'cat']
+
+da_cat_slot_values = {
+ # 'hotel-stay': ['1', '2', '3', '4', '5'],
+ 'hotel-internet': ['free', 'no', 'none', 'yes'],
+ 'hotel-parking': ['free', 'no', 'none', 'yes']
+}
+
+state_cat_slot_values = {}
+
+multiwoz_desc = {
+ 'taxi': {
+ 'domain': 'taxi information query system',
+ 'taxi_phone': 'taxi phone number',
+ 'taxi_types': 'taxi type',
+ },
+ 'restaurant': {
+ 'domain': 'restaurant information query system',
+ 'address': 'exact location of the restaurant',
+ 'postcode': 'postcode of the restaurant',
+ 'phone': 'restaurant phone number',
+ 'choice': 'number of restaurants meeting requests of user',
+ },
+ 'attraction': {
+ 'domain': 'an entertainment that is offered to the public',
+ 'address': 'details of where the attraction is',
+ 'postcode': 'postcode of the attraction',
+ 'phone': 'phone number of the attraction',
+ 'entrance fee': 'the fee charged for admission to the attraction',
+ 'pricerange': 'the price range for the attraction, from cheap to expensive',
+ 'choice': 'number of attractions matching requests of user'
+ },
+ 'booking': {
+ 'domain': 'to arrange with a taxi, restaurant, train, etc.',
+ 'time': 'time for an order',
+ 'day': 'day for an order, from monday to sunday',
+ 'stay': 'for how long the user wish to be at a place',
+ 'people': 'how many person the order is for',
+ 'name': 'name of the ordered place',
+ 'Ref': 'reference number of the order'
+ },
+ 'train': {
+ 'domain': 'query and order a train',
+ 'duration': 'the length of time the train trip lasts',
+ 'Ref': 'reference number of the order',
+ 'price': 'price for the train ticket',
+ 'choice': 'number of trains that meets requests of the user',
+ },
+ 'hotel': {
+ 'domain': 'to query hotel information and place an order',
+ 'address': 'exact location of the hotel',
+ 'postcode': 'postcode of the hotel',
+ 'phone': 'hotel phone number',
+ 'choice': 'number of hotels that meets requests of the user',
+ },
+ 'police': {
+ 'domain': 'find police stations',
+ 'address': 'exact location of the police station',
+ 'postcode': 'postcode of the police station',
+ 'phone': 'police station phone number',
+ },
+ 'intents': {
+ 'inform': 'inform user of value for a certain slot',
+ 'request': 'ask for value of a slot',
+ 'nobook': 'inform user of booking failure',
+ 'reqmore': 'ask user for more instructions',
+ 'book': 'place an order for user',
+ 'bye': 'end a conversation and say goodbye to user',
+ 'thank': 'express gratitude',
+ 'welcome': 'welcome',
+ 'offerbooked': 'inform user that an order is succussful',
+ 'recommend': 'recommend a choice for user request',
+ 'greet': 'express greeting',
+ 'nooffer': 'inform user that no options matches user request',
+ 'offerbook': 'offer to place an order for user',
+ 'select': 'provide several choices for user to choose from',
+ }
+}
+
+digit2word = {
+ '0': 'zero', '1': 'one', '2': 'two', '3': 'three', '4': 'four', '5': 'five',
+ '6': 'six', '7': 'seven', '8': 'eight', '9': 'nine', '10': 'ten'
+}
+
+
+def pharse_in_sen(phrase, sen):
+ '''
+ match value in the sentence
+ :param phrase: str
+ :param sen: str
+ :return: start, end if matched, else None, None
+ '''
+ assert isinstance(phrase, str)
+ pw = '(^|[\s,\.:\?!-])(?P<v>{})([\s,\.:\?!-]|$)'
+ pn = '(^|[\s\?!-]|\D[,\.:])(?P<v>{})($|[\s\?!-]|[,\.:]\D|[,\.:]$)'
+
+ if phrase.isdigit() and phrase in digit2word:
+ phrase = digit2word[phrase]
+ p = re.compile(pw.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ # match['0'] += 1
+ if phrase.isdigit():
+ pattern = pn
+ else:
+ pattern = pw
+ p = re.compile(pattern.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ return (None, None), 0
+
+
+
+
+
+def update_state(state, update):
+ # print('======================')
+ # print(state)
+ # print(update)
+ # print('======================')
+
+ for service, service_update in update.items():
+ if service not in state:
+ state[service] = copy.deepcopy(service_update)
+ else:
+ state[service].update(update[service])
+
+
+def convert_da(utt, da_dict, binary_ont, intent_ont, did, tid, da_cat_slot_values):
+ '''
+ convert multiwoz dialogue acts to required format
+ :param utt: user or system utt
+ :param da_dict: multiwoz da
+ :param binary_ont: binary ontology
+ :param intent_ont: intent ontology
+ :return:
+ '''
+ converted_da = {
+ 'categorical': [],
+ 'non-categorical': [],
+ 'binary': []
+ }
+
+ for Domain_Act, S, v in da_dict:
+ Domain, Act = Domain_Act.split('-')
+ if Domain.lower() in ['police', 'hospital', 'bus']:
+ continue
+
+ if Act.lower() not in intent_ont:
+ intent_ont[Act.lower()] = {}
+
+ # general domain is converted to empty domain. e.g. thank, bye
+ if Domain == 'general':
+ assert S == 'none'
+ assert v == 'none'
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': '',
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+
+
+ try:
+ reformated_slot = REF_SYS_DA[Domain][S]
+ except:
+ # print('44444444444444444444444444444444')
+ # print(Domain, S)
+ # logging.info('slot not in REF_SYS_DA, drop')
+ continue
+
+ # if slot is None, da should be converted into binary
+ if reformated_slot is None:
+ if not (S == 'none' and v == 'none'):
+ # mainly for `Open` slot
+ # print('11111111111111111111')
+ # print(Domain_Act, S, v)
+ continue
+ # Booking-Inform none none
+ # Police-Inform none none
+ # Train-OfferBook none none
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ reformated_domain_slot = Domain.lower() + '-' + reformated_slot
+
+ if Act.lower() == 'request':
+ converted_dict = {
+ 'intent': 'request',
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ # vs = da_dict[(Domain_Act, S)]['values']
+
+ if reformated_domain_slot in slot_to_type and slot_to_type[reformated_domain_slot] == 'cat':
+ origin_v = v
+ v = v.lower()
+ # if reformated_domain_slot in cat_slot_proj:
+ # v = cat_slot_proj[reformated_domain_slot][v]
+ if reformated_domain_slot not in da_cat_slot_values:
+ da_cat_slot_values[reformated_domain_slot] = []
+ # if v not in cat_slot_values[reformated_domain_slot]:
+ da_cat_slot_values[reformated_domain_slot].append(v)
+ converted_da['categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ if 'start_word' in da_dict[(Domain_Act, S, origin_v)]:
+ start_ws = da_dict[(Domain_Act, S, origin_v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, origin_v)]['end_word']
+ utt_list = utt.split()
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w)
+ if not utt[start_ch: end_ch] == origin_v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ converted_da['categorical'][-1].update({
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break
+
+ else:
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ # todo no span annotation
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ continue
+
+ start_ws = da_dict[(Domain_Act, S, v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, v)]['end_word']
+ utt_list = utt.split()
+ found = True
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w, v)
+ if not utt[start_ch: end_ch] == v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ found = True
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break
+
+ if not found:
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ return converted_da
+
+
+def get_state_update(prev_state, cur_state, dialog, did, tid, utt, coref_dict, slot_notfound_dict, da_cat_slot_values):
+ prev_turns = dialog['turns']
+ state_update = {'categorical': [], 'non-categorical': []}
+ notfoundnum = 0
+ total_value = 0
+
+ diff_state = {}
+ if prev_state is None:
+ diff_state = {domain: {slot: value for slot, value in cur_state[domain].items() if value != ''} for domain in
+ cur_state}
+ else:
+ assert len(prev_state) == len(cur_state), print(prev_state, cur_state)
+ for domain, domain_state in prev_state.items():
+ if domain not in diff_state:
+ diff_state[domain] = {}
+ for slot, value in domain_state.items():
+ if value != cur_state[domain][slot]:
+ # assert len(cur_state[domain][slot]) > 0, print(did, tid, domain, slot, utt)
+ diff_state[domain][slot] = cur_state[domain][slot]
+
+ ret_diff_state = copy.deepcopy(diff_state)
+
+
+
+ for domain in diff_state:
+ for slot in diff_state[domain]:
+
+ total_value += 1
+ fix_or = False
+ if '|' in diff_state[domain][slot]:
+ value = diff_state[domain][slot].split('|')[0]
+ else:
+ value = diff_state[domain][slot]
+
+ # if dialog['original_id'] == 'PMUL2512' and tid == 17 and value == '02:45':
+ # value = '2:45'
+
+ value_list = [value]
+ for _synonyms in synonyms:
+ if value in _synonyms:
+ value_list = _synonyms
+
+ value_list.extend(get_time_variants(value))
+ value_list.extend(get_genitive_variants(value))
+ value_list.extend(get_bb_variants(value))
+
+ if value.endswith(' restaurant'):
+ value_list.append(value.split(' restaurant')[0])
+ if value.endswith(' hotel'):
+ value_list.append(value.split(' hotel')[0])
+ found = False
+ for value in value_list:
+ # categorical slots
+ if slot in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']:
+ reformated_domain_slot = '-'.join([domain, slot])
+ if reformated_domain_slot in state_cat_slot_value_dict and (value in state_cat_slot_value_dict[reformated_domain_slot] or value in ['dontcare', '', 'none', 'not mentioned']):
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = [diff_state[domain][slot]]
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ else :
+ for _turn in prev_turns[::-1]:
+ found = False
+ for da in _turn['dialogue_act']['categorical']:
+ if da['value'] == value:
+ if 'start' in da:
+ state_update['categorical'][-1].update({
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ found = True
+ break
+ if found:
+ break
+ else:
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = [diff_state[domain][slot]]
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ # reformated_domain_slot = '-'.join([domain, slot]
+ found = True
+ break
+
+ # process value ---> none
+ assert value not in ['none', 'not mentioned']
+ if value in ['', 'dontcare']:
+ # if reformated_domain_slot not in state_cat_slot_values:
+ # state_cat_slot_values[reformated_domain_slot] = []
+ # # if v not in cat_slot_values[reformated_domain_slot]:
+ # state_cat_slot_values[reformated_domain_slot].append(value)
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ found = True
+ break
+
+ # first look for values in coref_dict
+ for _Domain_Act, _Slot, _value in coref_dict:
+ _domain, _act = _Domain_Act.lower().split('-')
+ _slot = _Slot.lower()
+ _coref_value = coref_dict[(_Domain_Act, _Slot, _value)]['coref_value']
+ if _coref_value == '':
+ continue
+ _coref_turn = coref_dict[(_Domain_Act, _Slot, _value)]['turn']
+ if _coref_turn == -1:
+ continue
+ _coref_pos = coref_dict[(_Domain_Act, _Slot, _value)]['pos']
+ if _coref_pos == '':
+ continue
+ _utt = coref_dict[(_Domain_Act, _Slot, _value)]['utt']
+ if _domain == domain and _slot == slot and value == _coref_value:
+
+ start_w, end_w = [int(p) for p in _coref_pos.split('-')]
+ utt_list = _utt.split()
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w + 1):
+ end_ch += len(utt_list[i]) + 1
+ end_ch -= 1
+
+ if not _utt[start_ch: end_ch] == _coref_value:
+ # print(111111111111111111111111111111111)
+ # print(_utt[start_ch: end_ch], _coref_value)
+ continue
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'from': 'coref',
+ 'utt_idx': _coref_turn,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ found = True
+
+ if found:
+ break
+
+ # from da annotation
+ for _turn in prev_turns[::-1]:
+ for da in _turn['dialogue_act']['non-categorical']:
+ # if da['domain'] == domain and da['slot'] == slot and fuzz.ratio(da['value'], value) > 85:
+ # if not da['value'] == value:
+ # print(1111111111111111)
+ # print(value, da['value'])
+
+ if fuzz.ratio(da['value'], value) > 85:
+
+ if 'start' in da:
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ # 'value': da['value'],
+ 'value': diff_state[domain][slot],
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if da['value'] != value:
+ state_update['non-categorical'][-1].update({'fixed_value':da['value']})
+ ret_diff_state[domain][slot] = da['value']
+
+ break
+ if found:
+ break
+
+ if found:
+ break
+
+ # from utterance
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+ (start, end), num = pharse_in_sen(str(value), _utt)
+ if num:
+ assert value.lower() == _utt[start:end].lower() \
+ or digit2word[value].lower() == _utt[start:end].lower()
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': _utt[start:end].lower(),
+ # 'fixed_value': _utt[start:end].lower(),
+ 'from': 'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': start,
+ 'end': end
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if value != _utt[start:end].lower():
+ state_update['non-categorical'][-1].update({'fixed_value': _utt[start:end].lower()})
+ ret_diff_state[domain][slot] = _utt[start:end].lower()
+ found = True
+ break
+ if found:
+ break
+
+ # from utterance
+ if not value.isdigit():
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+
+ s = difflib.SequenceMatcher(None, _utt, value)
+ matches = s.get_matching_blocks()
+
+ for i, j, n in matches:
+ possible_value = _utt[i: i+len(value)]
+
+ if i+ len(value) < len(_utt) and _utt[i+len(value)] not in [ ' ', ',', '.', '?', '!', '/'] :
+ possible_value += _utt[i+len(value):].split()[0]
+
+ if possible_value.startswith('th '):
+ possible_value = possible_value[3:]
+ i += 3
+ if i > 0 and _utt[i-1] not in [ ' ', ',', '.', '?', '!', '/']:
+ # cut first incomplete word
+ if len(possible_value.split()) > 1:
+ i += len(possible_value.split()[0]) + 1
+ possible_value = ' '.join(possible_value.split()[1:])
+
+
+ # prepend first incomplete word
+ # possible_value = _utt[:i].split()[-1] + possible_value
+ # i -= len(_utt[:i].split()[-1])
+
+
+ if fuzz.token_sort_ratio(value, possible_value) > 92 or possible_value.startswith('ashley hotel and lovell lodge') :
+ found = True
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': possible_value,
+ # 'fixed_value': possible_value,
+ 'from':'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': i,
+ 'end': i+len(possible_value)
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if possible_value != value:
+ state_update['non-categorical'][-1].update({'fixed_value': possible_value})
+ ret_diff_state[domain][slot] = possible_value
+ break
+ # assert _utt[i:i+len(possible_value)] == possible_value, print(_utt, _utt[i:i+len(possible_value)], possible_value)
+ # break
+ # if not possible_value == value:
+ # print(3333333333333333)
+ # print(value)
+ # print(possible_value)
+ if found:
+ break
+ if found:
+ break
+
+ if found:
+ break
+ if not found:
+ # print('3333333333333333333')
+ # print(did, tid)
+ # print(domain, slot, value)
+ # print([_t['utterance'] for _t in prev_turns])
+ # assert slot not in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']
+
+ if (domain, slot) not in slot_notfound_dict:
+ slot_notfound_dict[(domain, slot)] = 1
+ else:
+ slot_notfound_dict[(domain, slot)] += 1
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ return state_update, notfoundnum, total_value, ret_diff_state
+
+
+def merge_data_annotation():
+ extract_dir = os.path.join(self_dir, 'original_data')
+ data21 = json.load(open(os.path.join(self_dir, extract_dir, 'data.json')))
+ # data21_train = json.load(open(os.path.join(self_dir, extract_dir, 'train.json')))
+ # data21_val = json.load(open(os.path.join(self_dir, extract_dir, 'val.json')))
+ # data21_test = json.load(open(os.path.join(self_dir, extract_dir, 'test.json')))
+ # data21 = {}
+ # data21.update(data21_train)
+ # data21.update(data21_val)
+ # data21.update(data21_test)
+
+ # update_from_25_cnt = 0
+ # total_turn = 0
+ # for dial_id, dialog in data21.items():
+ # dial_id = dial_id + '.json'
+ # assert dial_id in data25
+ # for i, _turn in enumerate(dialog['log']):
+ # total_turn += 1
+ # if _turn['text'] == data25[dial_id]['log'][i]['text']:
+ # _turn['span_info'].extend(copy.deepcopy(data25[dial_id]['log'][i]['span_info']))
+ # # _turn['span_info'] = list(set(_turn['span_info']))
+ # # _turn['dialog_act'].update(copy.deepcopy(data25[dial_id]['log'][i]['dialog_act']))
+ # for Domain_Intent in data25[dial_id]['log'][i]['dialog_act']:
+ # if Domain_Intent in _turn['dialog_act']:
+ # _turn['dialog_act'][Domain_Intent].extend(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # else:
+ # _turn['dialog_act'][Domain_Intent] = copy.deepcopy(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # # _turn['dialog_act'][Domain_Intent] = list(set(_turn['dialog_act'][Domain_Intent]))
+ # if 'coreference' in data25[dial_id]['log'][i]:
+ # _turn['coreference'] = copy.deepcopy(data25[dial_id]['log'][i]['coreference'])
+ # update_from_25_cnt += 1
+ # else:
+ # # print('==============multiwoz21=================')
+ # # print(_turn['text'])
+ # # print('==============multiwoz25=================')
+ # # print(data25[dial_id]['log'][i]['text'])
+ # continue
+ #
+ # print('{}/{} turns update from multiwoz25 data'.format(update_from_25_cnt, total_turn))
+ return data21
+
+
+def preprocess(da_cat_slot_values, state_cat_slot_values):
+ all_data = []
+ binary_ont = []
+ intent_ont = {}
+ state_ont = {}
+
+ data_splits = ['train', 'val', 'test']
+ # data_splits = ['test']
+ extract_dir = os.path.join(self_dir, 'original_data')
+ num_train_dialogue = 0
+ num_train_utt = 0
+
+ num_match_error_da_span = 0
+
+ if not os.path.exists('data.zip') or not os.path.exists('ontology.json'):
+ # for data_split in data_splits:
+ # data_zip_file = os.path.join(self_dir, 'original_data.zip')
+ # if not os.path.exists(data_zip_file):
+ # raise FileNotFoundError(data_zip_file)
+
+ # logging.info('unzip multiwoz data to {}'.format(extract_dir))
+ # archive = zipfile.ZipFile(data_zip_file, 'r')
+ # archive.extractall(extract_dir)
+
+ data = merge_data_annotation()
+ # exit()
+ # data = json.load(open(os.path.join(self_dir, extract_dir, 'data_meta_fixed.json')))
+ train_list = open(os.path.join(self_dir, extract_dir, 'trainListFile')).read().split()
+ val_list = open(os.path.join(self_dir, extract_dir, 'valListFile')).read().split()
+ test_list = open(os.path.join(self_dir, extract_dir, 'testListFile')).read().split()
+
+ total_not_found_slot = 0
+ total_slot = 0
+ total_turn = 0
+ total_not_found_turn = 0
+ total_not_found_state = 0
+
+ slot_notfound_dict = {}
+
+ dialog_idx = 0
+ for dialog_id, dialog in tqdm(data.items()):
+
+ acc_not_found_flag = False
+
+ coref_dict = {}
+
+ data_split = None
+ for _split in data_splits:
+ if dialog_id.strip('.json') in eval(_split + '_list'):
+ data_split = _split
+ break
+ # assert data_split is not None
+ # if data_split != 'test':
+ # continue
+ if data_split is None:
+ continue
+
+ if data_split == 'train':
+ num_train_dialogue += len(data)
+
+ dialog_idx += 1
+ # if dialog_idx > 10:
+ # break
+ converted_dialogue = {
+ 'dataset': 'multiwoz22',
+ 'data_split': data_split,
+ 'dialogue_id': 'multiwoz22_' + str(dialog_idx),
+ 'original_id': dialog_id,
+ 'domains': [d for d in dialog['goal'] if
+ len(dialog['goal'][d]) != 0 and d in multiwoz_desc and d not in ['police', 'hospital', 'bus']],
+ 'turns': [],
+ }
+
+ if data_split == 'train':
+ num_train_utt += len(dialog['log'])
+
+ prev_state = None
+ accum_fixed_state = {}
+ for turn_id, turn in enumerate(dialog['log']):
+
+ utt = turn['text'].lower()
+ # for several wrong words
+ utt = utt.replace('seeuni', 'see uni')
+
+ utt = ' '.join(utt.split())
+ utt = utt.replace(' im ', ' i\'m ')
+ utt = utt.replace(' dont ', ' don\'t ')
+ utt = utt.replace(' thats ', ' that\'s ')
+ utt = utt.replace('idon\'t', ' i don\'t ')
+ utt = utt.replace('wedon\'t ', 'we don\'t ')
+ utt = utt.replace('id be ', 'i\'d be ')
+ # utt = utt.replace('cambridgethat\'svery ', 'cambridge that\'s very')
+ utt = re.sub(r'^im ', 'i\'m ', utt)
+ utt = re.sub(r'^whats ', 'what\'s ', utt)
+ utt = re.sub(r'^id ', 'i\'d ', utt)
+ utt = re.sub(r'^thats ', 'that\'s ', utt)
+
+ utt = re.sub( r'([a-zA-Z0-9])([,.!\'-\?"~])', r'\1 \2', utt)
+ utt = re.sub(r'([,.!\'-\?"~])([a-zA-Z0-9])', r'\1 \2', utt)
+
+ das = turn.get('dialog_act', [])
+ role = 'user' if turn_id % 2 == 0 else 'system'
+ spans = turn.get('span_info', [])
+
+ da_dict = {}
+ for Domain_Act in das:
+ Domain = Domain_Act.split('-')[0]
+ if Domain.lower() not in converted_dialogue['domains'] and Domain.lower() not in ['general', 'booking']:
+ continue
+
+ Svs = das[Domain_Act]
+ for S, v in Svs:
+ v = v.lower()
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+
+ if (Domain_Act, S, v) not in da_dict:
+ da_dict[(Domain_Act, S, v)] = {}
+
+ for span in spans:
+ Domain_Act, S, v, start_word, end_word = span
+ v = v.lower()
+ if not (Domain_Act, S, v) in da_dict:
+ # logging.info('span da annotation not found in multiwoz da label')
+ # logging.info(dialog_id, turn_id)
+ # logging.info((Domain_Act, S, v))
+ # logging.info(da_dict)
+ num_match_error_da_span += 1
+ else:
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ da_dict[(Domain_Act, S, v)]['start_word'] = []
+ da_dict[(Domain_Act, S, v)]['end_word'] = []
+
+ da_dict[(Domain_Act, S, v)]['start_word'].append(start_word)
+ da_dict[(Domain_Act, S, v)]['end_word'].append(end_word)
+
+ converted_turn = {
+ 'utt_idx': turn_id,
+ 'speaker': role,
+ 'utterance': utt,
+ 'dialogue_act': convert_da(utt, da_dict, binary_ont, intent_ont, dialog_id, turn_id, da_cat_slot_values),
+ }
+
+ # for state annotations
+ if role == 'system':
+ turn_state = turn['metadata']
+ cur_state = {}
+ for domain in turn_state:
+ if domain in ['police', 'hospital', 'bus']:
+ continue
+ if domain not in converted_dialogue['domains']:
+ continue
+ cur_state[domain] = {}
+ for subdomain in ['semi', 'book']:
+ for slot in turn_state[domain][subdomain]:
+ if slot == 'booked':
+ continue
+ if slot == 'ticket': # or (domain == 'train' and slot == 'people'):
+ # for cases where domain slot exists in REF but not in state
+ # because of check in evaluate.py
+ continue
+
+ else:
+ fixed_slot = slot
+ state_ds = domain + '-' + fixed_slot
+ if state_ds not in slot_to_type:
+ logging.info('state slot not defined in da list')
+ logging.info(state_ds)
+ if turn_state[domain][subdomain][slot] in ['', [], 'not mentioned', 'none']:
+ cur_state[domain][fixed_slot] = ""
+ else:
+ if turn_state[domain][subdomain][slot].startswith('th '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+ if turn_state[domain][subdomain][slot].startswith('he '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+
+ cur_state[domain][fixed_slot] = turn_state[domain][subdomain][slot]
+
+ if domain not in state_ont:
+ state_ont[domain] = []
+ if fixed_slot not in state_ont[domain]:
+ state_ont[domain].append(fixed_slot)
+
+ if domain == 'train' and 'people' not in cur_state[domain]:
+ cur_state[domain]['people'] = ''
+ # if len(converted_turn['state'][domain]) == 0:
+ # converted_turn['state'].pop(domain)
+ if len(converted_dialogue['turns']) > 0:
+ # move state from system side to user side
+ converted_dialogue['turns'][-1]['state'] = copy.deepcopy(cur_state)
+
+ # for state update annotations
+ state_update, _notfoundslot, _totalslot, ret_diff_state = get_state_update(prev_state, cur_state, converted_dialogue,
+ dialog_id, turn_id, turn['text'], coref_dict,
+ slot_notfound_dict, da_cat_slot_values)
+
+ update_state(accum_fixed_state, ret_diff_state)
+ for domain in accum_fixed_state:
+ for slot in accum_fixed_state[domain]:
+ assert isinstance(accum_fixed_state[domain][slot], str), print(accum_fixed_state[domain][slot])
+
+ if _notfoundslot == 0:
+ # for slot in state_update['categorical']:
+ # assert 'fixed_value' not in slot
+ for slot in state_update['non-categorical']:
+ if slot['value'] not in ['', 'dontcare']:
+ assert 'utt_idx' in slot
+
+ else:
+ flag = False
+ for slot in state_update['categorical']:
+ if 'fixed_value' in slot:
+ flag = True
+ break
+ for slot in state_update['non-categorical']:
+ if 'utt_idx' not in slot:
+ flag = True
+ break
+ assert flag, print(flag, state_update['non-categorical'])
+
+ total_turn += 1
+ total_slot += _totalslot
+ total_not_found_slot += _notfoundslot
+ total_not_found_turn += 1 if _notfoundslot > 0 else 0
+ if _notfoundslot > 0:
+ acc_not_found_flag = True
+ if acc_not_found_flag:
+ total_not_found_state += 1
+
+ coref_dict = {}
+ converted_dialogue['turns'][-1]['state_update'] = copy.deepcopy(state_update)
+ converted_dialogue['turns'][-1]['fixed_state'] = copy.deepcopy(accum_fixed_state)
+ if 'state' not in converted_dialogue['turns'][-1]:
+ converted_dialogue['turns'][-1]['state'] = {}
+ prev_state = copy.deepcopy(cur_state)
+
+ converted_dialogue['turns'].append(converted_turn)
+
+ if 'coreference' in turn:
+ for Domain_Act in turn['coreference']:
+ for Slot, value, coref, coref_turn, coref_pos in turn['coreference'][Domain_Act]:
+ value = value.lower()
+ coref_dict[(Domain_Act, Slot, value)] = {'turn': coref_turn, 'pos': coref_pos,
+ 'coref_value': coref,
+ 'utt': converted_dialogue['turns'][coref_turn][
+ 'utterance']}
+
+ check_spans(converted_dialogue)
+ # postprocess_update_spans(converted_dialogue)
+ if converted_dialogue['turns'][-1]['speaker'] == 'system':
+ converted_dialogue['turns'].pop(-1)
+ all_data.append(converted_dialogue)
+
+ print('total_turn', total_turn)
+ print('total_not_found_turn', total_not_found_turn)
+ print('total_slot', total_slot)
+ print('total_not_found_slot', total_not_found_slot)
+ print('total_not_found_state', total_not_found_state)
+ print(slot_notfound_dict)
+ from collections import Counter
+ # print({k : dict(Counter(v)) for k, v in cat_slot_values.items()})
+ json.dump({k : dict(Counter(v)) for k, v in state_cat_slot_values.items()}, open(os.path.join(self_dir, 'cat_slot_values.json'), 'w'), indent=4)
+ cat_slot_values = {k: list(set(v)) for k, v in state_cat_slot_values.items()}
+ da_cat_slot_values = {k: list(set(v)) for k, v in da_cat_slot_values.items()}
+
+ json.dump(all_data, open('data.json', 'w'), indent=4)
+ write_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ os.remove('data.json')
+
+ new_ont = {
+ 'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': {}
+ }
+
+ for d_s in slot_to_type:
+ d, s = d_s.split('-')
+ if d not in new_ont['domains']:
+ new_ont['domains'][d] = {
+ 'description': multiwoz_desc[d]['domain'],
+ 'slots': {}
+ }
+ domain_ont = new_ont['domains'][d]
+ assert s not in domain_ont
+ domain_ont['slots'][s] = {
+ 'description': multiwoz_desc[d][s] if s in multiwoz_desc[d] else '',
+ 'is_categorical': d_s in state_cat_slot_ds,
+ 'possible_values': da_cat_slot_values[d_s] if d_s in state_cat_slot_ds else []
+ }
+ domain_ont['slots'][s]['possible_values'] = [_ for _ in domain_ont['slots'][s]['possible_values'] if _ not in ['dontcare', '']]
+
+ new_ont['state'] = {}
+ # print(state_cat_slot_value_dict)
+ print(state_ont)
+ for d in state_ont:
+ new_ont['state'][d] = {}
+ for s in state_ont[d]:
+ d_s = '-'.join([d, s])
+ new_ont['state'][d][s] = ''
+
+ new_ont['intents'] = {i: {'description': multiwoz_desc['intents'][i]} for i in intent_ont}
+ new_ont['binary_dialogue_act'] = binary_ont
+
+ slot_desc = json.load(open(os.path.join(self_dir, extract_dir, './slot_descriptions.json')))
+ for domain_slot in slot_desc:
+ _domain, _slot = domain_slot.split('-')
+ _desc = slot_desc[domain_slot][0]
+ if _slot == 'arriveby':
+ _slot = 'arriveBy'
+ elif _slot == 'leaveat':
+ _slot = 'leaveAt'
+ if 'book' in _slot:
+ _slot = _slot.replace('book ', '')
+ if not _domain in new_ont['state']:
+ # logging.info('domain {} not in state domains'.format(_domain))
+ continue
+ if _domain in new_ont['domains'] and _slot in new_ont['domains'][_domain]['slots']:
+ new_ont['domains'][_domain]['slots'][_slot]['description'] = _desc
+ if not _slot in new_ont['state'][_domain]:
+ logging.info('domain {} slot {} not in state'.format(_domain, _slot))
+ continue
+ # new_ont['state'][_domain][_slot] = ""
+ assert _domain in new_ont['domains'], print(_domain)
+ assert _slot in new_ont['domains'][_domain]['slots']
+
+ logging.info('num_match_error_da_span {}'.format(num_match_error_da_span))
+ json.dump(new_ont, open(os.path.join(self_dir, './ontology.json'), 'w'), indent=4)
+
+ else:
+ all_data = read_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ new_ont = json.load(open(os.path.join(self_dir, './ontology.json'), 'r'))
+ logging.info('# dialogue: {}, # turn: {}'.format(num_train_dialogue, num_train_utt))
+ return all_data, new_ont
+
+
+# def postprocess_update_spans(dialog):
+# changed_utt_idx_and_position = {}
+# for turn in dialog['turns']:
+# if turn['speaker'] != 'user':
+# continue
+# changed = False
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+#
+# # assume at most one word changes for every utterance
+# if turn['utt_idx'] not in changed_utt_idx_and_position:
+# if utt_idx == turn['utt_idx'] and start-1 > -1 and turn['utterance'][start-1] not in [' ']:
+# changed_utt_idx_and_position[turn['utt_idx']] = start
+# print('=======================')
+# print(dialog['original_id'])
+# print(turn['utterance'])
+# print(json.dumps(_update, indent=2))
+# print(turn['utterance'][start: end])
+# turn['utterance'] = turn['utterance'][:start] + ' ' + turn['utterance'][start:]
+# print(turn['utterance'])
+# _update['start'] += 1
+# _update['end'] += 1
+# changed = True
+# if utt_idx not in changed_utt_idx_and_position:
+# continue
+# else:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update['fixed_value'] != 'not found' else _update['value']
+# if start >= changed_utt_idx_and_position[utt_idx]:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+# _update['start'] += 1
+# _update['end'] += 1
+# elif start < changed_utt_idx_and_position[utt_idx] < end:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1], value)
+# print('fix')
+# print(_update)
+# _update['end'] += 1
+# _update['fixed_value'] = turn['utterance'][_update['start']: _update['end'] + 1].strip()
+# print(_update)
+# if changed:
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+#
+# if utt_idx not in changed_utt_idx_and_position:
+# continue
+# else:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+# 'fixed_value'] != 'not found' else _update['value']
+# if start >= changed_utt_idx_and_position[utt_idx]:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# assert dialog['turns'][utt_idx]['utterance'][_update['start'] + 1: _update['end'] + 1] == value
+# _update['start'] += 1
+# _update['end'] += 1
+# elif start < changed_utt_idx_and_position[utt_idx] < end:
+# if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+# print('====================fix===================')
+# print(_update)
+# assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+# _update['end'] += 1
+# _update['fixed_value'] = dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end'] + 1]
+# print(_update)
+# for turn in dialog['turns']:
+# if turn['speaker'] != 'user':
+# continue
+# for _update in turn['state_update']['non-categorical']:
+# if 'utt_idx' in _update:
+# value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+# 'fixed_value'] != 'not found' else _update['value']
+# utt_idx = _update['utt_idx']
+# start = _update['start']
+# end = _update['end']
+# if dialog['turns'][utt_idx]['utterance'][start] == ' ':
+# _update['start'] += 1
+# _update['fixed_value'] = value[1:]
+# value = value[1:]
+# start += 1
+# assert dialog['turns'][utt_idx]['utterance'][start: end] == value, print(json.dumps(turn, indent=4), [c for c in dialog['turns'][utt_idx]['utterance'][start: end]], [c for c in value])
+# return dialog
+
+
+def get_time_variants(time_text):
+ value_list = [time_text]
+ pattern_time = r'(\d{1,2}:\d{2})(\s)?(am|pm|AM|PM)?'
+ match_times = re.findall(pattern_time, time_text)
+ if len(match_times) < 1:
+ return []
+ match_time = match_times[0]
+
+ am_flag = match_time[2] in ['am', 'AM']
+ pm_flag = match_time[2] in ['pm', 'PM']
+ no_am_pm_flag = match_time[2] == ''
+ if am_flag:
+ # 4:00am -> 4:00
+ value_list.append(match_time[0])
+ if len(match_time[0]) == 4:
+ # 4:00 -> 04:00
+ value_list.append('0' + match_time[0])
+ if pm_flag:
+ # 4:00pm -> 16:00
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ new_hour = 12 + hour
+ value_list.append(str(new_hour)+':'+min)
+ if no_am_pm_flag:
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ if hour > 12:
+ new_hour = hour - 12
+ value_list.append(str(new_hour) + ':' + min + 'pm')
+ value_list.append(str(new_hour) + ':' + min + ' pm')
+ value_list.append(str(new_hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(new_hour) + 'pm')
+ value_list.append(str(new_hour) + ' pm')
+ value_list.append(str(new_hour))
+ else:
+ value_list.append(str(hour) + ':' + min + 'am')
+ value_list.append(str(hour) + ':' + min + ' am')
+ value_list.append(str(hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(hour) + 'am')
+ value_list.append(str(hour) + ' am')
+ value_list.append(str(hour))
+ if len(match_time[0]) == 5 and match_time[0][0] == '0':
+ value_list.append(match_time[0][1:])
+ value_list.append(''.join(match_time[0].split(':')))
+
+ return value_list
+
+
+def get_genitive_variants(value):
+ ret_list = []
+ value_genitive_format = r"(?=\w)s(?=\s)"
+ value_pattern = re.compile(value_genitive_format)
+
+ span_genitive_value = re.sub(value_pattern, " 's", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ span_genitive_value = re.sub(value_pattern, "'s", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ # if len(ret_list) > 0:
+ # print('=============================')
+ # print(value)
+ # print(re.findall(value_pattern, value))
+ # print(ret_list)
+ return ret_list
+
+
+def check_spans(dialog):
+ for turn in dialog['turns']:
+ if turn['speaker'] != 'user':
+ continue
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+ 'fixed_value'] != 'not found' else _update['value']
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+ assert dialog['turns'][utt_idx]['utterance'][start:end] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][start:end])
+
+
+
+def get_bb_variants(value):
+ ret_list = []
+ if 'bed and breakfast' in value:
+ ret_list.append(value.replace('bed and breakfast', 'b & b'))
+ return ret_list
+
+if __name__ == '__main__':
+ preprocess(da_cat_slot_values, state_cat_slot_values)
\ No newline at end of file
diff --git a/data/unified_datasets/multiwoz23/README.md b/data/unified_datasets/multiwoz23/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..57b5a077f7dc3c467764a664b6d01d08414c6e23
--- /dev/null
+++ b/data/unified_datasets/multiwoz23/README.md
@@ -0,0 +1,33 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots. state and state updates.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 8434 | 105066 | 12.46 | 15.75 | 7 |
+| dev | 999 | 13731 | 13.74 | 16.1 | 7 |
+| train | 1000 | 13744 | 13.74 | 16.08 | 7 |
+
+
+## Main changes
+
+- only keep 5 domains in state annotations and dialog acts.
+- `pricerange`, `area`, `day`, `internet`, `parking`, `stars` are considered categorical slots.
+- replace special tokens by space. e.g. `I want@to find a hotel. -> I want to find a hotel.`
+
+Run `evaluate.py`:
+
+da values match rate: 98.798
+state values match rate: 89.185
+
+### original data
+
+- from [multiwoz-coref](https://github.com/lexmen318/MultiWOZ-coref) repo.
+- slot description by multiwoz2.2
+- some hand-written descriptions.
+
+
diff --git a/data/unified_datasets/multiwoz23/data.zip b/data/unified_datasets/multiwoz23/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..54d08b07f6ca6f70fdfe39f9bdfda3d4564c1e16
Binary files /dev/null and b/data/unified_datasets/multiwoz23/data.zip differ
diff --git a/data/unified_datasets/multiwoz23/ontology.json b/data/unified_datasets/multiwoz23/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..3d2c0a8dcc659e287ee88c95cf65ee9f7b3b3e8c
--- /dev/null
+++ b/data/unified_datasets/multiwoz23/ontology.json
@@ -0,0 +1,1983 @@
+{
+ "domains": {
+ "taxi": {
+ "description": "taxi information query system",
+ "slots": {
+ "destination": {
+ "description": "destination of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of taxi",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_phone": {
+ "description": "taxi phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "taxi_types": {
+ "description": "taxi type",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "restaurant": {
+ "description": "restaurant information query system",
+ "slots": {
+ "food": {
+ "description": "food type for the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget for the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "budget conscious",
+ "same area and price range",
+ "expensive side",
+ "mostly expensive",
+ "fairly expensive",
+ "all price ranges",
+ "that range",
+ "vietnamese",
+ "same area and price range as the hotel",
+ "middle",
+ "or otherwise",
+ "inexpensive",
+ "on the pricey side",
+ "ranging from cheap to expensive",
+ "moderatly priced",
+ "cherap",
+ "centre",
+ "more moderate",
+ "north",
+ "quite low",
+ "this price range",
+ "same price range",
+ "moderatley priced",
+ "varying price range",
+ "great prices",
+ "fine",
+ "expensive price",
+ "more expensive",
+ "upscale",
+ "any price range",
+ "cheapish",
+ "moderatre",
+ "modest",
+ "pricey",
+ "same price",
+ "fairly cheap",
+ "relatively cheap",
+ "cheap priced",
+ "moderate pricing",
+ "expensive price range",
+ "cheaply priced",
+ "east",
+ "expensive range",
+ "quite inexpensive",
+ "same area and price",
+ "moderately prices",
+ "or the moderately priced",
+ "expensive but worth every penny",
+ "assorted",
+ "any",
+ "any price",
+ "on the cheap side",
+ "do n't care",
+ "budget friendly",
+ "most affordable price range",
+ "low cost",
+ "moderate|cheap",
+ "pretty cheap",
+ "a little expensive but worth it",
+ "epensive",
+ "not too pricey",
+ "moderate priced",
+ "not - so - expensive",
+ "no particular",
+ "your",
+ "cheaper",
+ "priced cheaply",
+ "expensive priced",
+ "fairly inexpensive",
+ "moderate and one in the cheap range",
+ "reasonable",
+ "expensive(no",
+ "cheap price range",
+ "expinsive",
+ "mostly expensive and moderately priced",
+ "quite expensive",
+ "very expensive",
+ "that price",
+ "cheap range",
+ "boderate",
+ "reasonably priced",
+ "rather expensive",
+ "any price point",
+ "moderate or expensive priced",
+ "all",
+ "luxury options",
+ "regardless of price",
+ "expensive / upscale",
+ "moderate price",
+ "similar price range",
+ "pretty expensive",
+ "same area and pricerange",
+ "that price range",
+ "moderate range",
+ "moderately expensive",
+ "moderately",
+ "not",
+ "moderatly",
+ "cheaply",
+ "does n't matter",
+ "budget - friendly",
+ "cheapt",
+ "cheapest",
+ "various",
+ "not moderate",
+ "expensive or moderate",
+ "not cheap",
+ "expensively priced",
+ "expensive and moderately priced",
+ "not really",
+ "high class",
+ "low priced",
+ "your price range",
+ "varying price",
+ "do not care",
+ "not too expensive",
+ "adforable",
+ "they vary",
+ "none",
+ "moderate price range",
+ "high - end",
+ "moderately price",
+ "not expensive",
+ "relatively expensive",
+ "cheap to expensive",
+ "affordable",
+ "cheaply - priced",
+ "expensive",
+ "moderate",
+ "moderately - priced",
+ "priced moderately",
+ "same area and price range as my hotel",
+ "moderate and expensive",
+ "less expensive",
+ "expensively",
+ "high end",
+ "most expensive",
+ "same pricerange",
+ "does not matter",
+ "do not have a preference",
+ "cheap or expensive",
+ "the least expensive",
+ "and",
+ "moderately priced",
+ "cheap",
+ "every price point",
+ "the same"
+ ]
+ },
+ "address": {
+ "description": "exact location of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "east area of town",
+ "the city centre",
+ "that part of town",
+ "centre city",
+ "west of town",
+ "center",
+ "close to the hotel",
+ "in cambridge",
+ "ctre",
+ "south side of the town",
+ "close to your location",
+ "town centre area",
+ "any area",
+ "east side of the city",
+ "that area of town",
+ "west part of town .",
+ "on the east",
+ "north of town",
+ "n the centre",
+ "that area",
+ "east part of town",
+ "the west of town",
+ "centre",
+ "west of cambridge",
+ "north",
+ "south area",
+ "all of cambridge",
+ "centre of town",
+ "same area as the park",
+ "centrally located",
+ "the north",
+ "these areas",
+ "all of",
+ "north or centre",
+ "downtown",
+ "centre area",
+ "westside",
+ "the east part of town",
+ "throughout the area",
+ "near clare hall",
+ "the south",
+ "the east",
+ "the north area",
+ "the south side",
+ "east section of town",
+ "by the airport",
+ "the south area of the city",
+ "the west side of town",
+ "the south part of town",
+ "centre region",
+ "centre area of the city",
+ "close it city centre",
+ "mexican",
+ "centre of the city",
+ "anywhere in the city",
+ "central",
+ "north side of town",
+ "near the center of town",
+ "north and west",
+ "east",
+ "the south area of town",
+ "here",
+ "west area",
+ "in town at all",
+ "the east side",
+ "centrally",
+ "the centre area of town",
+ "around town",
+ "east side",
+ "centre of cambridge",
+ "in the north",
+ "centreof",
+ "central region of town",
+ "any",
+ "the west area",
+ "the west part of town",
+ "west part of the city",
+ "same side",
+ "ely",
+ "other part of the town",
+ "the north part of town",
+ "do n't care",
+ "same area",
+ "anywhere in town",
+ "same price range and area",
+ "north end",
+ "the north end",
+ "here in the city",
+ "same side of town as your hotel",
+ "the area you have chosen",
+ "center of the town",
+ "the north side of town",
+ "near the centre",
+ "the same area as the botanic gardens",
+ "southern",
+ "north part of town",
+ "this town",
+ "central area",
+ "east of cambridge",
+ "east part of tow",
+ "do n't have a preference",
+ "northern parts of town",
+ "south side of town",
+ "centre cambridge",
+ "anywhere",
+ "south part of time",
+ "north area of town",
+ "the center",
+ "all over the city",
+ "any where in the city",
+ "the town centre",
+ "southern area",
+ "the north side .",
+ "north area",
+ "in the city .",
+ "in the city centre",
+ "the city center",
+ "south side",
+ "east end",
+ "near centre",
+ "same part of town",
+ "anywhere in cambridge",
+ "town center",
+ "the center of town",
+ "south of cambridge",
+ "west side of town",
+ "that",
+ "thai",
+ "in the west",
+ "south area of town",
+ "northern part",
+ "does",
+ "the east side of town",
+ "northern part of cambridge",
+ "same location",
+ "centre area of the town",
+ "same",
+ "southend",
+ "center of town",
+ "the south of town",
+ "north side",
+ "the centre area",
+ "south of town .",
+ "west cambridge",
+ "the are",
+ "in town",
+ "do nt care",
+ "cambridge centre",
+ "not",
+ "in the city",
+ "the east area",
+ "city",
+ "the centre of town",
+ "does n't matter",
+ "the centre",
+ "the area",
+ "centre of town .",
+ "cetre",
+ "towns centre",
+ "west end of town",
+ "around the college",
+ "the west side",
+ "town",
+ "west",
+ "wet part of town",
+ "eastside",
+ "centrem",
+ "the southern area",
+ "north cambridge",
+ "in the east",
+ "east area",
+ "the south side of town",
+ "westies",
+ "chesterton",
+ "elsewhere in the city",
+ "not in the west",
+ "any part of town",
+ "other parts of town",
+ "the same area",
+ "town of centre",
+ "centere",
+ "same areas",
+ "town centre",
+ "east side of town",
+ "close to that area",
+ "none",
+ "center cambridge",
+ "city center",
+ "southside",
+ "same side of town",
+ "not in the south of town",
+ "north side of chersteron",
+ "around there",
+ "west part of town",
+ "the north side",
+ "south end",
+ "the area of west",
+ "northern",
+ "centre part of town",
+ "center area of town",
+ "city centre",
+ "restaurants in the city",
+ "west area of town",
+ "expensive",
+ "all over cambridge",
+ "south part of the city",
+ "east section",
+ "this area",
+ "within this area",
+ "eastern",
+ "nearby",
+ "west side area",
+ "centre-",
+ "the west",
+ "south part",
+ "throughout the city",
+ "south of town",
+ "same area as the hotel",
+ "closely located",
+ "south cambridge",
+ "and centre",
+ "the center part of town",
+ "west side",
+ "does not matter",
+ "east of town",
+ "almost every area of town",
+ "north side of cambridge",
+ "south part of town",
+ "all over town",
+ "same area of town",
+ "northside",
+ "cambridge",
+ "south",
+ "the west end",
+ "cheap",
+ "the centre part of town",
+ "in the centre",
+ "centre area of town",
+ "east|south",
+ "the same",
+ "west end"
+ ]
+ },
+ "postcode": {
+ "description": "postcode of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "restaurant phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "time of the restaurant booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the restaurant booking",
+ "is_categorical": true,
+ "possible_values": [
+ "sunday|thursday",
+ "sunday",
+ "friday",
+ "saturday",
+ "saturday|thursday",
+ "wednesday",
+ "monday",
+ "thursday",
+ "same day",
+ "tuesday",
+ "same group and day"
+ ]
+ },
+ "choice": {
+ "description": "number of restaurants meeting requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "attraction": {
+ "description": "an entertainment that is offered to the public",
+ "slots": {
+ "address": {
+ "description": "details of where the attraction is",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "phone number of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "entrance fee": {
+ "description": "the fee charged for admission to the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "type of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "area": {
+ "description": "area or place of the attraction",
+ "is_categorical": true,
+ "possible_values": [
+ "east area of town",
+ "close to the center of town",
+ "the city centre",
+ "every area except the north",
+ "that part of town",
+ "centre city",
+ "center",
+ "center area",
+ "in cambridge",
+ "near ely",
+ "all of the other areas",
+ "town centre area",
+ "south part of cambridge",
+ "the east of the town",
+ "the center of the park",
+ "center of the city",
+ "west of your city",
+ "the area you 're looking for",
+ "north of town",
+ "cenre",
+ "that area",
+ "east part of town",
+ "the west of town",
+ "centre area of cambridge",
+ "centre",
+ "north section of cambridge",
+ "north",
+ "the south area",
+ "south area",
+ "near the hotel",
+ "west area of the city",
+ "centre of town",
+ "the north of the city",
+ "centrally located",
+ "the north",
+ "central district",
+ "northern cambridge",
+ "city 's centre",
+ "the center area",
+ "southern cambridge",
+ "northend",
+ "downtown",
+ "centre area",
+ "westside",
+ "the east part of town",
+ "north of the city",
+ "the south",
+ "the east",
+ "museums",
+ "the south side",
+ "the west side of town",
+ "centre region",
+ "centre of the city",
+ "east end of town",
+ "church area",
+ "central",
+ "east",
+ "west area",
+ "the west area of town",
+ "the east side",
+ "same general area",
+ "city cenre",
+ "east side",
+ "centre of cambridge",
+ "centre|west",
+ "any",
+ "west part of the city",
+ "the west area",
+ "same side",
+ "the eastside",
+ "all around the city",
+ "do n't care",
+ "same area",
+ "in the south",
+ "there",
+ "the area you are looking for",
+ "close to the restaurant",
+ "another area",
+ "center of the town",
+ "very close in the same area",
+ "near the centre",
+ "central zone",
+ "southern",
+ "north part of town",
+ "does n't really matter",
+ "east of cambridge",
+ "do n't have a preference",
+ "that side of town",
+ "western part of the town",
+ "the south near your hotel",
+ "north in milton",
+ "south side of town",
+ "eat",
+ "cetnre of town",
+ "the center",
+ "all over the city",
+ "cambridge leisure park",
+ "north area",
+ "most of them are in the center",
+ "different area",
+ "south side",
+ "that area .",
+ "museum",
+ "town center",
+ "same part of town",
+ "western cambridge",
+ "south of cambridge",
+ "west side of town",
+ "that",
+ "in that side",
+ "center of cambridge",
+ "does",
+ "whole of cambridge",
+ "all",
+ "to the south",
+ "same location",
+ "same",
+ "center part of town",
+ "center of town",
+ "norwich",
+ "city centre .",
+ "north side",
+ "the centre area",
+ "the centry area",
+ "west cambridge",
+ "west end of the city",
+ "your desired location",
+ "in town",
+ "entre",
+ "in the center",
+ "not",
+ "centre of town ?",
+ "city",
+ "on the centre",
+ "the centre of town",
+ "does n't matter",
+ "went side of town",
+ "the centre",
+ "various",
+ "the area",
+ "towns centre",
+ "right in the center of town",
+ "in the area",
+ "west end of town",
+ "central cambridge",
+ "the west side",
+ "eastside",
+ "west",
+ "this side of town",
+ "west side of the city",
+ "near the restaurant",
+ "cb30aq",
+ "center of town .",
+ "not really",
+ "in the east",
+ "same area as tandoori palace",
+ "west part of cambridge",
+ "centre by galleria",
+ "east area",
+ "wet end",
+ "do not care",
+ "western",
+ "northern area of town",
+ "western area of town",
+ "other parts of town",
+ "the same area",
+ "town centre",
+ "the same road",
+ "east side of town",
+ "center are",
+ "none",
+ "city center",
+ "centre area .",
+ "in that area",
+ "clifton way",
+ "west part of town",
+ "south end",
+ "western part of town",
+ "several different parts of town",
+ "northern",
+ "centre part of town",
+ "city centre",
+ "west area of town",
+ "east section",
+ "this area",
+ "nearby",
+ "the west",
+ "same part of town as your restaurant",
+ "throughout the city",
+ "south park of town",
+ "south of town",
+ "other areas",
+ "south cambridge",
+ "the west - side",
+ "the center part of town",
+ "west side",
+ "does not matter",
+ "east of town",
+ "do not have a preference",
+ "same area of town",
+ "south part of town",
+ "all over town",
+ "cambridge",
+ "south",
+ "the west end",
+ "same part",
+ "the centre part of town",
+ "in the centre",
+ "centre area of town",
+ "east cambridge",
+ "northern area",
+ "cambridge 's centre",
+ "west end"
+ ]
+ },
+ "name": {
+ "description": "name of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "the price range for the attraction, from cheap to expensive",
+ "is_categorical": true,
+ "possible_values": [
+ "5 pound entrance fee",
+ "unaware of what their entrance fee is",
+ "2 pounds",
+ "we do n't have any information",
+ "4 pounds",
+ "no entrance fee",
+ "do n't have information",
+ "5 pounds",
+ "do n't see a price listed",
+ "expensive",
+ "neither prices are listed",
+ "not listed",
+ "entrance fee",
+ "free",
+ "free admission",
+ "3.50 pounds",
+ "not sure of the fee",
+ "moderately priced",
+ "cheap",
+ "from free to 2 pounds 50",
+ "they do n't have the entrance fee posted"
+ ]
+ },
+ "choice": {
+ "description": "number of attractions matching requests of user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "booking": {
+ "description": "to arrange with a taxi, restaurant, train, etc.",
+ "slots": {
+ "time": {
+ "description": "time for an order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day for an order, from monday to sunday",
+ "is_categorical": true,
+ "possible_values": [
+ "one",
+ "sundar",
+ "2",
+ "6",
+ "tonight",
+ "that time frame",
+ "that",
+ "this evening",
+ "saturday night thru monday",
+ "we d",
+ "sunday",
+ "wednesday",
+ "same day",
+ "at that time",
+ "tuesday",
+ "sunday and monday",
+ "friday",
+ "vmhj6y3i",
+ "1",
+ "monday",
+ "tues",
+ "the same day",
+ "sunday 18/06/2017",
+ "tuesday instead of thursday",
+ "vmhj6y3i.",
+ "weds",
+ "tuesday 's",
+ "saturday",
+ "today",
+ "thursday",
+ "tuesday through friday"
+ ]
+ },
+ "stay": {
+ "description": "for how long the user wish to be at a place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "how many person the order is for",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the ordered place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "train": {
+ "description": "query and order a train",
+ "slots": {
+ "destination": {
+ "description": "destination of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "arriveBy": {
+ "description": "arrival time of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure": {
+ "description": "departure location of the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "duration": {
+ "description": "the length of time the train trip lasts",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "people": {
+ "description": "number of people booking for train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the train",
+ "is_categorical": true,
+ "possible_values": [
+ "everday",
+ "this day",
+ "every day",
+ "13:07",
+ "sundays",
+ "none",
+ "that",
+ "thursdays",
+ "sunday",
+ "frday",
+ "wednesday",
+ "same day",
+ "between friday and wednesday",
+ "tuesday",
+ "tr2519",
+ "sat",
+ "all week",
+ "friday",
+ "saturdays",
+ "monday",
+ "weds",
+ "that day",
+ "every",
+ "saturday",
+ "other days are available",
+ "thursday",
+ "fiday",
+ "cambridge",
+ "daily",
+ "saturday-",
+ "mondays"
+ ]
+ },
+ "Ref": {
+ "description": "reference number of the order",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaveAt": {
+ "description": "leaving time for the train",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "price for the train ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "choice": {
+ "description": "number of trains that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel": {
+ "description": "to query hotel information and place an order",
+ "slots": {
+ "internet": {
+ "description": "internet option at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "no",
+ "yes",
+ "none",
+ "free"
+ ]
+ },
+ "area": {
+ "description": "area or place of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "east area of town",
+ "north location",
+ "that part of town",
+ "centre city",
+ "west of town",
+ "center",
+ "east and the north .",
+ "in cambridge",
+ "southern part of town",
+ "the city 's south side",
+ "other parts of the city",
+ "close to where you 'll be dining",
+ "east near other shops and boutiques",
+ "any area",
+ "that area of town",
+ "same area as the restaurant",
+ "north end of the city",
+ "centrally - located",
+ "north of town",
+ "northern side of town",
+ "that area",
+ "east part of town",
+ "centre",
+ "on the west",
+ "north",
+ "the south area",
+ "south area",
+ "all of cambridge",
+ "eastern part of the city",
+ "centre of town",
+ "east part of time",
+ "the north of the city",
+ "centrally located",
+ "on the south",
+ "the north",
+ "northern cambridge",
+ "all over",
+ "northend",
+ "downtown",
+ "centre area",
+ "westside",
+ "northern part of town",
+ "least",
+ "the east",
+ "the south side",
+ "east section of town",
+ "the south",
+ "south end of the city",
+ "across town",
+ "centre of the city",
+ "or west ?",
+ "centra",
+ "south closer to the museum",
+ "north side of town",
+ "east",
+ "el shaddai",
+ "different part of town",
+ "west area",
+ "all areas",
+ "on the west side",
+ "nearby the restaurant",
+ "the east side",
+ "same general area",
+ "you are interested in",
+ "centrally",
+ "north park of town",
+ "east side",
+ "in the north",
+ "centre of cambridge",
+ "not to far from the restaurant",
+ "any",
+ "west part of the city",
+ "the west area",
+ "centre area of town",
+ "centre part",
+ "that vicinity",
+ "the north part of town",
+ "all around the city",
+ "do n't care",
+ "same area",
+ "the north end",
+ "north part of the city",
+ "north end",
+ "there",
+ "and in the centre",
+ "north and centre",
+ "center of the town",
+ "near the centre",
+ "west side near the restaurant",
+ "that region",
+ "southern",
+ "north part of town",
+ "except in the east",
+ "another area of town",
+ "near the museum",
+ "nborth",
+ "do n't have a preference",
+ "no particular",
+ "that side of town",
+ "south side of town",
+ "centre cambridge",
+ "eat",
+ "north area of town",
+ "the",
+ "all over the city",
+ "north par of town",
+ "north part of town .",
+ "next door",
+ "the town centre",
+ "different area",
+ "north area",
+ "south side",
+ "any part of the city",
+ "near centre",
+ "west part",
+ "everywhere but the city centre",
+ "same part of town",
+ "anywhere in cambridge",
+ "western cambridge",
+ "the center of town",
+ "north of cambridge",
+ "west side of town",
+ "in the west",
+ "south area of town",
+ "the east side of town",
+ "wast",
+ "southend",
+ "center of town",
+ "west areas of town",
+ "north side",
+ "on the eastside",
+ "the centre area",
+ "west end of the city",
+ "central location",
+ "in town",
+ "in the center",
+ "not",
+ "the centre of the city",
+ "in the city",
+ "the east area",
+ "the easy",
+ "the centre of cambridge",
+ "the centre of town",
+ "does n't matter",
+ "various parts of the city",
+ "several areas of town",
+ "norht",
+ "the centre",
+ "various",
+ "the area",
+ "not in the south",
+ "west|centre",
+ "in the area",
+ "on the north",
+ "the west side",
+ "eastside",
+ "west",
+ "near the restaurant",
+ "north cambridge",
+ "not really",
+ "south cambridge area",
+ "throughout cambridge",
+ "any part of town",
+ "western",
+ "other parts of town",
+ "the same area",
+ "town centre",
+ "east side of town",
+ "none",
+ "do n't really care",
+ "city center",
+ "southside",
+ "and east",
+ "all across town",
+ "south end of town",
+ "different parts of the city",
+ "same side of town",
+ "west part of town",
+ "the north side",
+ "south end",
+ "except in the north",
+ "northern",
+ "centre part of town",
+ "city centre",
+ "eastern cambridge",
+ "the town center",
+ "eastern",
+ "the east end of town",
+ "all parts of the city",
+ "different parts of town",
+ "the west",
+ "throughout the city",
+ "north end of town",
+ "south of town",
+ "a different area",
+ "south cambridge",
+ "the northern part of town",
+ "west side",
+ "across cambridge",
+ "does not matter",
+ "east of town",
+ "do not have a preference",
+ "all over town",
+ "south part of town",
+ "same area of town",
+ "somewhere else",
+ "northside",
+ "the westside",
+ "cambridge",
+ "south",
+ "the west end",
+ "northe part of town",
+ "cheap",
+ "in the centre",
+ "east area",
+ "east cambridge",
+ "northern area",
+ "the same",
+ "west end"
+ ]
+ },
+ "stars": {
+ "description": "star rating of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "4|5",
+ "one",
+ "2",
+ "four - star",
+ "unrated",
+ "does not show",
+ "3|4",
+ "none",
+ "different star ratings",
+ "four",
+ "4",
+ "not as fancy",
+ "3",
+ "yes",
+ "drop the star rating",
+ "four starts",
+ "no star rating",
+ "3-star",
+ "0 to 4",
+ "5",
+ "2-stars",
+ "does not have",
+ "1",
+ "two",
+ "lower",
+ "2-star",
+ "1-star",
+ "4-stars",
+ "ranging from 2 - 4 stars",
+ "three",
+ "2-starred",
+ "foru",
+ "0-star",
+ "3-stars",
+ "0",
+ "4-star",
+ "no",
+ "several",
+ "zero",
+ "five",
+ "5-star",
+ "not rated"
+ ]
+ },
+ "parking": {
+ "description": "parking facility at the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "no",
+ "yes",
+ "none",
+ "free"
+ ]
+ },
+ "phone": {
+ "description": "hotel phone number",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "name": {
+ "description": "name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pricerange": {
+ "description": "price budget of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "moderately pricing",
+ "same area and price range",
+ "different price ranges",
+ "that range",
+ "moderate to cheap range",
+ "cheaply prices",
+ "or expensive ?",
+ "all different price ranges",
+ "inexpensive",
+ "cheap|moderate",
+ "affordable pricing",
+ "is moderately priced",
+ "your chosen",
+ "moderatly priced",
+ "different",
+ "the cheapest",
+ "centre",
+ "moderate prices",
+ "same part of town and price range as the restaurant",
+ "cheaper side",
+ "north",
+ "moderately priced .",
+ "more budget - friendly",
+ "lower",
+ "cheap > moderate",
+ "same price range",
+ "moderatley priced",
+ "cheap>moderate",
+ "more expensive",
+ "upscale",
+ "free",
+ "different price range",
+ "very affordable",
+ "moderate to cheap",
+ "moderately to expensively priced",
+ "modrate",
+ "same price",
+ "fairly cheap",
+ "super cheap",
+ "relatively cheap",
+ "cheap priced",
+ "expensive price range",
+ "moderate pricing",
+ "your price",
+ "cheaply priced",
+ "expensive range",
+ "same area and price range as the restaurant",
+ "same area and price range as my restaurant",
+ "moderately prices",
+ "even cheaper",
+ "moderate or cheap",
+ "moderately - priced or cheap",
+ "higher price range",
+ "any",
+ "oderately priced",
+ "other ranges",
+ "lower end",
+ "great",
+ "extremely reasonable",
+ "on the cheap side",
+ "budget - priced",
+ "do n't care",
+ "low cost",
+ "varying price ranges",
+ "moderate|cheap",
+ "pretty cheap",
+ "quite cheap",
+ "moderate priced",
+ "cheap to moderate",
+ "no particular",
+ "cheaper",
+ "very inexpensive",
+ "does not say",
+ "epxensive",
+ "expensively - priced",
+ "cheap price range",
+ "expensive or cheap",
+ "on the more expensive side",
+ "more moderately priced",
+ "hotel",
+ "moderate in price",
+ "quite expensive",
+ "cheap range",
+ "that",
+ "slightly more expensive",
+ "moderately priceed",
+ "rather expensive",
+ "all",
+ "does",
+ "moderate price",
+ "chear",
+ "pretty expensive",
+ "moderatly price",
+ "that price range",
+ "moderate range",
+ "much cheaper",
+ "a little pricey",
+ "cheap side",
+ "moderately",
+ "not",
+ "mostly in the cheap to moderate price range",
+ "economically priced",
+ "cheaply",
+ "does n't matter",
+ "cheapest",
+ "range",
+ "fairly cheap compared to other hotels",
+ "expensively priced",
+ "very cheap",
+ "not really",
+ "your price range",
+ "do not care",
+ "none",
+ "less costly",
+ "moderate price range",
+ "inexpensively - priced",
+ "moderately price",
+ "$ 100",
+ "espensive",
+ "not expensive",
+ "cheap to expensive",
+ "cheap or moderate",
+ "cheaply - priced",
+ "affordable",
+ "moderate price point",
+ "guesthouses",
+ "expensive",
+ "priced moderately",
+ "moderate",
+ "moderately - priced",
+ "same general price",
+ "cheaper than the others",
+ "moderately pried",
+ "expensive to moderate",
+ "unfortunately do not have the price",
+ "ca n't view that information",
+ "$100",
+ "hotels",
+ "same pricerange",
+ "does not matter",
+ "moderately priced",
+ "economical",
+ "cheap",
+ "the same"
+ ]
+ },
+ "people": {
+ "description": "number of people for the hotel booking",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "exact location of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "what is the type of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "stay": {
+ "description": "length of stay at the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "day": {
+ "description": "day of the hotel booking",
+ "is_categorical": true,
+ "possible_values": [
+ "thursday--",
+ "wednesday|friday",
+ "none",
+ "friday > tuesday",
+ "sunday",
+ "same",
+ "wednesday",
+ "same day",
+ "tuesday",
+ "monday < thursday",
+ "friday",
+ "saturday|tuesday",
+ "friday>tuesday",
+ "monday",
+ "monday<thursday",
+ "tuesday/",
+ "saturday",
+ "thursday",
+ "sunday>monday",
+ "sunday > monday"
+ ]
+ },
+ "choice": {
+ "description": "number of hotels that meets requests of the user",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "police": {
+ "description": "find police stations",
+ "slots": {
+ "address": {
+ "description": "exact location of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "postcode": {
+ "description": "postcode of the police station",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone": {
+ "description": "police station phone number",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "inform user of value for a certain slot"
+ },
+ "request": {
+ "description": "ask for value of a slot"
+ },
+ "nobook": {
+ "description": "inform user of booking failure"
+ },
+ "reqmore": {
+ "description": "ask user for more instructions"
+ },
+ "book": {
+ "description": "place an order for user"
+ },
+ "bye": {
+ "description": "end a conversation and say goodbye to user"
+ },
+ "thank": {
+ "description": "express gratitude"
+ },
+ "welcome": {
+ "description": "welcome"
+ },
+ "offerbook": {
+ "description": "offer to place an order for user"
+ },
+ "offerbooked": {
+ "description": "inform user that an order is succussful"
+ },
+ "recommend": {
+ "description": "recommend a choice for user request"
+ },
+ "greet": {
+ "description": "express greeting"
+ },
+ "nooffer": {
+ "description": "inform user that no options matches user request"
+ },
+ "select": {
+ "description": "provide several choices for user to choose from"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "stay",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "reqmore",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "bye",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "thank",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "welcome",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "day",
+ "value": ""
+ },
+ {
+ "intent": "offerbook",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "greet",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "internet",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "parking",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "food",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "Ref",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "nobook",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "people",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "stars",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "entrance fee",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "booking",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "pricerange",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "attraction",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform",
+ "domain": "taxi",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "departure",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_types",
+ "value": ""
+ },
+ {
+ "intent": "book",
+ "domain": "booking",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "taxi_phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "arriveBy",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "taxi",
+ "slot": "leaveAt",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "nooffer",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "hotel",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "attraction",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "name",
+ "value": ""
+ },
+ {
+ "intent": "offerbooked",
+ "domain": "train",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "recommend",
+ "domain": "restaurant",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "state": {
+ "hotel": {
+ "name": "",
+ "area": "",
+ "parking": "",
+ "pricerange": "",
+ "stars": "",
+ "internet": "",
+ "type": "",
+ "stay": "",
+ "day": "",
+ "people": ""
+ },
+ "train": {
+ "leaveAt": "",
+ "destination": "",
+ "day": "",
+ "arriveBy": "",
+ "departure": "",
+ "people": ""
+ },
+ "restaurant": {
+ "food": "",
+ "pricerange": "",
+ "name": "",
+ "area": "",
+ "time": "",
+ "day": "",
+ "people": ""
+ },
+ "attraction": {
+ "type": "",
+ "name": "",
+ "area": ""
+ },
+ "taxi": {
+ "leaveAt": "",
+ "destination": "",
+ "departure": "",
+ "arriveBy": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/multiwoz23/original_data.zip b/data/unified_datasets/multiwoz23/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..68bffe4bfedf2ecc860c5e204323a91e8a6882c8
Binary files /dev/null and b/data/unified_datasets/multiwoz23/original_data.zip differ
diff --git a/data/unified_datasets/multiwoz23/preprocess.py b/data/unified_datasets/multiwoz23/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..d293e7f9c0f2c78f60750f858a34e76d2100ecb4
--- /dev/null
+++ b/data/unified_datasets/multiwoz23/preprocess.py
@@ -0,0 +1,1421 @@
+import copy
+import re
+import zipfile
+import json
+import os
+from tqdm import tqdm
+import sys
+import difflib
+from fuzzywuzzy import fuzz
+sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
+
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+import logging
+
+
+logging.basicConfig(level=logging.INFO)
+self_dir = (os.path.abspath(os.getcwd()))
+
+REF_SYS_DA = {
+ 'Attraction': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Fee': "entrance fee", 'Name': "name", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange", 'Type': "type",
+ 'none': None, 'Open': None
+ },
+ 'Hospital': {
+ 'Department': 'department', 'Addr': 'address', 'Post': 'postcode',
+ 'Phone': 'phone', 'none': None
+ },
+ 'Booking': {
+ 'Day': 'day', 'Name': 'name', 'People': 'people',
+ 'Ref': 'Ref', 'Stay': 'stay', 'Time': 'time',
+ 'none': None
+ },
+ 'Hotel': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Internet': "internet", 'Name': "name", 'Parking': "parking",
+ 'Phone': "phone", 'Post': "postcode", 'Price': "pricerange",
+ 'Stars': "stars", 'Type': "type", 'Stay': 'stay', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Restaurant': {
+ 'Addr': "address", 'Area': "area", 'Choice': "choice",
+ 'Name': "name", 'Food': "food", 'Phone': "phone",
+ 'Post': "postcode", 'Price': "pricerange",
+ 'Time': 'time', 'Day': 'day', 'People': 'people',
+ 'none': None
+ },
+ 'Taxi': {
+ 'Arrive': "arriveBy", 'Car': "taxi_types", 'Depart': "departure",
+ 'Dest': "destination", 'Leave': "leaveAt", 'Phone': "taxi_phone",
+ 'none': None
+ },
+ 'Train': {
+ 'Arrive': "arriveBy", 'Choice': "choice", 'Day': "day",
+ 'Depart': "departure", 'Dest': "destination",
+ 'Leave': "leaveAt", 'People': "people", 'Ref': "Ref",
+ 'Time': "duration", 'none': None, 'Ticket': 'price',
+ },
+ 'Police': {
+ 'Addr': "address", 'Post': "postcode", 'Phone': "phone", 'none': None
+ },
+}
+
+# taxi restaurant attraction train
+slot_to_type = {
+ 'taxi-destination': 'non',
+ 'taxi-departure': 'non',
+ 'taxi-leaveAt': 'non',
+ 'taxi-arriveBy': 'non',
+ 'restaurant-food': 'non',
+ 'restaurant-name': 'non',
+ 'attraction-address': 'non',
+ 'attraction-postcode': 'non',
+ 'restaurant-pricerange': 'cat',
+ 'restaurant-address': 'non',
+ 'restaurant-area': 'cat',
+ 'restaurant-postcode': 'non',
+ 'attraction-phone': 'non',
+ 'attraction-entrance fee': 'non',
+ 'booking-time': 'non',
+ 'booking-day': 'cat',
+ 'attraction-type': 'non',
+ 'attraction-area': 'cat',
+ 'train-destination': 'non',
+ 'train-arriveBy': 'non',
+ 'train-departure': 'non',
+ 'hotel-internet': 'cat',
+ 'hotel-area': 'cat',
+ 'booking-stay': 'non',
+ 'booking-people': 'non',
+ 'train-duration': 'non',
+ 'train-people': 'non',
+ 'train-day': 'cat',
+ 'train-Ref': 'non',
+ 'hotel-stars': 'cat',
+ 'train-leaveAt': 'non',
+ 'train-price': 'non',
+ 'hotel-parking': 'cat',
+ 'hotel-phone': 'non',
+ 'hotel-name': 'non',
+ 'hotel-pricerange': 'cat',
+ 'hotel-people': 'non',
+ 'restaurant-phone': 'non',
+ 'hotel-postcode': 'non',
+ 'hotel-address': 'non',
+ 'attraction-name': 'non',
+ 'hotel-type': 'non',
+ 'restaurant-people': 'non',
+ 'train-choice': 'non',
+ 'attraction-pricerange': 'cat',
+ 'hotel-stay': 'non',
+ 'booking-name': 'non',
+ 'booking-Ref': 'non',
+ 'restaurant-time': 'non',
+ 'restaurant-day': 'cat',
+ 'hotel-day': 'cat',
+ 'hotel-choice': 'non',
+ 'restaurant-choice': 'non',
+ 'attraction-choice': 'non',
+ 'taxi-taxi_phone': 'non',
+ 'taxi-taxi_types': 'non',
+ 'police-address': 'non',
+ 'police-postcode': 'non',
+ 'police-phone': 'non'
+}
+
+state_cat_slot_value_dict = {
+ "hotel-pricerange": {
+ "cheap": 735,
+ "moderate": 1063,
+ "expensive": 594,
+ },
+ "hotel-parking": {
+ "yes": 1809,
+ "no": 126,
+ "free": 4,
+ },
+ "hotel-day": {
+ "tuesday": 385,
+ "wednesday": 410,
+ "monday": 365,
+ "saturday": 407,
+ "friday": 393,
+ "thursday": 384,
+ "sunday": 369,
+ },
+ "train-day": {
+ "wednesday": 533,
+ "monday": 533,
+ "saturday": 543,
+ "thursday": 547,
+ "friday": 563,
+ "tuesday": 553,
+ "sunday": 613,
+ },
+ "hotel-stars": {
+ "4": 1263,
+ "2": 193,
+ "0": 201,
+ "3": 401,
+ "5": 45,
+ "1": 45,
+ },
+ "hotel-internet": {
+ "yes": 1841,
+ "no": 79,
+ "free": 2
+ },
+ "hotel-area": {
+ "east": 416,
+ "north": 717,
+ "centre": 538,
+ "south": 289,
+ "west": 316,
+ },
+ "attraction-area": {
+ "centre": 1290,
+ "west": 332,
+ "north": 155,
+ "south": 240,
+ "east": 272,
+ },
+ "restaurant-pricerange": {
+ "expensive": 1477,
+ "cheap": 758,
+ "moderate": 1028,
+ },
+ "restaurant-area": {
+ "centre": 1745,
+ "south": 398,
+ "north": 390,
+ "east": 360,
+ "west": 423,
+ },
+ "restaurant-day": {
+ "thursday": 362,
+ "wednesday": 412,
+ "friday": 395,
+ "monday": 383,
+ "sunday": 399,
+ "saturday": 421,
+ "tuesday": 350,
+ }
+}
+
+
+synonyms = [
+ ["el shaddia guesthouse", "el shaddai"],
+ [ "peterborough", "peterbourgh"],
+ ["night club", "nightclub", 'nightclubs'],
+ ["boat", "boating"],
+ ["portugese", "portuguese"],
+ ["guesthouse", "guest house"],
+ ["seafood", "sea food"],
+ ["christ 's college", "christ college"],
+ ["huntingdon marriott hotel"]
+]
+
+state_cat_slot_ds = [k for k, v in slot_to_type.items() if v == 'cat']
+
+da_cat_slot_values = {
+ # 'hotel-stay': ['1', '2', '3', '4', '5'],
+ 'hotel-internet': ['free', 'no', 'none', 'yes'],
+ 'hotel-parking': ['free', 'no', 'none', 'yes'],
+}
+
+state_cat_slot_values = {}
+
+multiwoz_desc = {
+ 'taxi': {
+ 'domain': 'taxi information query system',
+ 'taxi_phone': 'taxi phone number',
+ 'taxi_types': 'taxi type',
+ },
+ 'restaurant': {
+ 'domain': 'restaurant information query system',
+ 'address': 'exact location of the restaurant',
+ 'postcode': 'postcode of the restaurant',
+ 'phone': 'restaurant phone number',
+ 'choice': 'number of restaurants meeting requests of user',
+ },
+ 'attraction': {
+ 'domain': 'an entertainment that is offered to the public',
+ 'address': 'details of where the attraction is',
+ 'postcode': 'postcode of the attraction',
+ 'phone': 'phone number of the attraction',
+ 'entrance fee': 'the fee charged for admission to the attraction',
+ 'pricerange': 'the price range for the attraction, from cheap to expensive',
+ 'choice': 'number of attractions matching requests of user'
+ },
+ 'booking': {
+ 'domain': 'to arrange with a taxi, restaurant, train, etc.',
+ 'time': 'time for an order',
+ 'day': 'day for an order, from monday to sunday',
+ 'stay': 'for how long the user wish to be at a place',
+ 'people': 'how many person the order is for',
+ 'name': 'name of the ordered place',
+ 'Ref': 'reference number of the order'
+ },
+ 'train': {
+ 'domain': 'query and order a train',
+ 'duration': 'the length of time the train trip lasts',
+ 'Ref': 'reference number of the order',
+ 'price': 'price for the train ticket',
+ 'choice': 'number of trains that meets requests of the user',
+ },
+ 'hotel': {
+ 'domain': 'to query hotel information and place an order',
+ 'address': 'exact location of the hotel',
+ 'postcode': 'postcode of the hotel',
+ 'phone': 'hotel phone number',
+ 'choice': 'number of hotels that meets requests of the user',
+ },
+ 'police': {
+ 'domain': 'find police stations',
+ 'address': 'exact location of the police station',
+ 'postcode': 'postcode of the police station',
+ 'phone': 'police station phone number',
+ },
+ 'intents': {
+ 'inform': 'inform user of value for a certain slot',
+ 'request': 'ask for value of a slot',
+ 'nobook': 'inform user of booking failure',
+ 'reqmore': 'ask user for more instructions',
+ 'book': 'place an order for user',
+ 'bye': 'end a conversation and say goodbye to user',
+ 'thank': 'express gratitude',
+ 'welcome': 'welcome',
+ 'offerbooked': 'inform user that an order is succussful',
+ 'recommend': 'recommend a choice for user request',
+ 'greet': 'express greeting',
+ 'nooffer': 'inform user that no options matches user request',
+ 'offerbook': 'offer to place an order for user',
+ 'select': 'provide several choices for user to choose from',
+ }
+}
+
+digit2word = {
+ '0': 'zero', '1': 'one', '2': 'two', '3': 'three', '4': 'four', '5': 'five',
+ '6': 'six', '7': 'seven', '8': 'eight', '9': 'nine', '10': 'ten'
+}
+
+
+def pharse_in_sen(phrase, sen):
+ '''
+ match value in the sentence
+ :param phrase: str
+ :param sen: str
+ :return: start, end if matched, else None, None
+ '''
+ assert isinstance(phrase, str)
+ pw = '(^|[\s,\.:\?!-])(?P<v>{})([\s,\.:\?!-]|$)'
+ pn = '(^|[\s\?!-]|\D[,\.:])(?P<v>{})($|[\s\?!-]|[,\.:]\D|[,\.:]$)'
+
+ if phrase.isdigit() and phrase in digit2word:
+ phrase = digit2word[phrase]
+ p = re.compile(pw.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ # match['0'] += 1
+ if phrase.isdigit():
+ pattern = pn
+ else:
+ pattern = pw
+ p = re.compile(pattern.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ # if num > 1:
+ # match['>1'] += 1
+ # else:
+ # match['1'] += 1
+ return m.span('v'), num
+ return (None, None), 0
+
+
+
+
+
+def update_state(state, update):
+ # print('======================')
+ # print(state)
+ # print(update)
+ # print('======================')
+
+ for service, service_update in update.items():
+ if service not in state:
+ state[service] = copy.deepcopy(service_update)
+ else:
+ state[service].update(update[service])
+
+
+def convert_da(utt, da_dict, binary_ont, intent_ont, did, tid, da_cat_slot_values):
+ '''
+ convert multiwoz dialogue acts to required format
+ :param utt: user or system utt
+ :param da_dict: multiwoz da
+ :param binary_ont: binary ontology
+ :param intent_ont: intent ontology
+ :return:
+ '''
+ converted_da = {
+ 'categorical': [],
+ 'non-categorical': [],
+ 'binary': []
+ }
+
+ for Domain_Act, S, v in da_dict:
+ Domain, Act = Domain_Act.split('-')
+ if Domain.lower() in ['police', 'hospital', 'bus']:
+ continue
+
+ if Act.lower() not in intent_ont:
+ intent_ont[Act.lower()] = {}
+
+ # general domain is converted to empty domain. e.g. thank, bye
+ if Domain == 'general':
+ assert S == 'none'
+ assert v == 'none'
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': '',
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+
+
+ try:
+ reformated_slot = REF_SYS_DA[Domain][S]
+ except:
+ # print('44444444444444444444444444444444')
+ # print(Domain, S)
+ # logging.info('slot not in REF_SYS_DA, drop')
+ continue
+
+ # if slot is None, da should be converted into binary
+ if reformated_slot is None:
+ if not (S == 'none' and v == 'none'):
+ # mainly for `Open` slot
+ # print('11111111111111111111')
+ # print(Domain_Act, S, v)
+ continue
+ # Booking-Inform none none
+ # Police-Inform none none
+ # Train-OfferBook none none
+ converted_dict = {
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': '',
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ reformated_domain_slot = Domain.lower() + '-' + reformated_slot
+
+ if Act.lower() == 'request':
+ converted_dict = {
+ 'intent': 'request',
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': ''
+ }
+ converted_da['binary'].append(converted_dict)
+
+ if converted_dict not in binary_ont:
+ binary_ont.append(converted_dict)
+ continue
+
+ # vs = da_dict[(Domain_Act, S)]['values']
+
+ if reformated_domain_slot in slot_to_type and slot_to_type[reformated_domain_slot] == 'cat':
+ origin_v = v
+ v = v.lower()
+ # if reformated_domain_slot in cat_slot_proj:
+ # v = cat_slot_proj[reformated_domain_slot][v]
+ if reformated_domain_slot not in da_cat_slot_values:
+ da_cat_slot_values[reformated_domain_slot] = []
+ # if v not in cat_slot_values[reformated_domain_slot]:
+ da_cat_slot_values[reformated_domain_slot].append(v)
+ converted_da['categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ if 'start_word' in da_dict[(Domain_Act, S, origin_v)]:
+ start_ws = da_dict[(Domain_Act, S, origin_v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, origin_v)]['end_word']
+ utt_list = utt.split()
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w)
+ if not utt[start_ch: end_ch] == origin_v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ converted_da['categorical'][-1].update({
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break;
+
+ else:
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ # todo no span annotation
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ continue
+
+ start_ws = da_dict[(Domain_Act, S, v)]['start_word']
+ end_ws = da_dict[(Domain_Act, S, v)]['end_word']
+ utt_list = utt.split()
+ found = True
+ for start_w, end_w in zip(start_ws, end_ws):
+ if start_w > len(utt_list) or end_w > len(utt_list):
+ continue
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w):
+ end_ch += len(utt_list[i]) + 1
+ try:
+ end_ch += len(utt_list[end_w])
+ except:
+ print(utt_list, start_w, end_w)
+ if not utt[start_ch: end_ch] == v:
+ # print('2222222222222222222222222')
+ # print('\n'.join([v, utt[start_ch: end_ch - 1]]))
+ continue
+
+ else:
+ found = True
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ break
+
+ if not found:
+ converted_da['non-categorical'].append({
+ 'intent': Act.lower(),
+ 'domain': Domain.lower(),
+ 'slot': reformated_slot,
+ 'value': v
+ })
+ return converted_da
+
+
+def get_state_update(prev_state, cur_state, dialog, did, tid, utt, coref_dict, slot_notfound_dict, da_cat_slot_values):
+ prev_turns = dialog['turns']
+ state_update = {'categorical': [], 'non-categorical': []}
+ notfoundnum = 0
+ total_value = 0
+
+ diff_state = {}
+ if prev_state is None:
+ diff_state = {domain: {slot: value for slot, value in cur_state[domain].items() if value != ''} for domain in
+ cur_state}
+ else:
+ assert len(prev_state) == len(cur_state), print(prev_state, cur_state)
+ for domain, domain_state in prev_state.items():
+ if domain not in diff_state:
+ diff_state[domain] = {}
+ for slot, value in domain_state.items():
+ if value != cur_state[domain][slot]:
+ # assert len(cur_state[domain][slot]) > 0, print(did, tid, domain, slot, utt)
+ diff_state[domain][slot] = cur_state[domain][slot]
+
+ ret_diff_state = copy.deepcopy(diff_state)
+
+
+
+ for domain in diff_state:
+ for slot in diff_state[domain]:
+
+ total_value += 1
+ fix_or = False
+ if '|' in diff_state[domain][slot]:
+ value = diff_state[domain][slot].split('|')[0]
+ else:
+ value = diff_state[domain][slot]
+
+ # if dialog['original_id'] == 'PMUL2512' and tid == 17 and value == '02:45':
+ # value = '2:45'
+
+ value_list = [value]
+ for _synonyms in synonyms:
+ if value in _synonyms:
+ value_list = _synonyms
+
+ value_list.extend(get_time_variants(value))
+ value_list.extend(get_genitive_variants(value))
+ value_list.extend(get_bb_variants(value))
+
+ if value.endswith(' restaurant'):
+ value_list.append(value.split(' restaurant')[0])
+ if value.endswith(' hotel'):
+ value_list.append(value.split(' hotel')[0])
+ found = False
+ for value in value_list:
+ # categorical slots
+ if slot in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']:
+ reformated_domain_slot = '-'.join([domain, slot])
+ if reformated_domain_slot in state_cat_slot_value_dict and (value in state_cat_slot_value_dict[reformated_domain_slot] or value in ['dontcare', '', 'none', 'not mentioned']):
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = [diff_state[domain][slot]]
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ else :
+ for _turn in prev_turns[::-1]:
+ found = False
+ for da in _turn['dialogue_act']['categorical']:
+ if da['value'] == value:
+ if 'start' in da:
+ state_update['categorical'][-1].update({
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ found = True
+ break
+ if found:
+ break
+ else:
+ state_update['categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ if domain + '-' + slot not in da_cat_slot_values:
+ da_cat_slot_values[domain + '-' + slot] = []
+ da_cat_slot_values[domain + '-' + slot].append(diff_state[domain][slot])
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ # reformated_domain_slot = '-'.join([domain, slot]
+ found = True
+ break
+
+ # process value ---> none
+ assert value not in ['none', 'not mentioned']
+ if value in ['', 'dontcare']:
+ # if reformated_domain_slot not in state_cat_slot_values:
+ # state_cat_slot_values[reformated_domain_slot] = []
+ # # if v not in cat_slot_values[reformated_domain_slot]:
+ # state_cat_slot_values[reformated_domain_slot].append(value)
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot]
+ })
+ found = True
+ break
+
+ # first look for values in coref_dict
+ for _Domain_Act, _Slot, _value in coref_dict:
+ _domain, _act = _Domain_Act.lower().split('-')
+ _slot = _Slot.lower()
+ _coref_value = coref_dict[(_Domain_Act, _Slot, _value)]['coref_value']
+ if _coref_value == '':
+ continue
+ _coref_turn = coref_dict[(_Domain_Act, _Slot, _value)]['turn']
+ if _coref_turn == -1:
+ continue
+ _coref_pos = coref_dict[(_Domain_Act, _Slot, _value)]['pos']
+ if _coref_pos == '':
+ continue
+ _utt = coref_dict[(_Domain_Act, _Slot, _value)]['utt']
+ if _domain == domain and _slot == slot and value == _coref_value:
+
+ start_w, end_w = [int(p) for p in _coref_pos.split('-')]
+ utt_list = _utt.split()
+ start_ch = 0
+ for i in range(start_w):
+ start_ch += len(utt_list[i]) + 1
+ end_ch = start_ch
+ for i in range(start_w, end_w + 1):
+ end_ch += len(utt_list[i]) + 1
+ end_ch -= 1
+
+ if not _utt[start_ch: end_ch] == _coref_value:
+ # print(111111111111111111111111111111111)
+ # print(_utt[start_ch: end_ch], _coref_value)
+ continue
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'from': 'coref',
+ 'utt_idx': _coref_turn,
+ 'start': start_ch,
+ 'end': end_ch
+ })
+ if value != diff_state[domain][slot]:
+ state_update['categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ found = True
+
+ if found:
+ break
+
+ # from da annotation
+ for _turn in prev_turns[::-1]:
+ for da in _turn['dialogue_act']['non-categorical']:
+ # if da['domain'] == domain and da['slot'] == slot and fuzz.ratio(da['value'], value) > 85:
+ # if not da['value'] == value:
+ # print(1111111111111111)
+ # print(value, da['value'])
+
+ if fuzz.ratio(da['value'], value) > 85:
+
+ if 'start' in da:
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ # 'value': da['value'],
+ 'value': diff_state[domain][slot],
+ 'utt_idx': _turn['utt_idx'],
+ 'start': da['start'],
+ 'end': da['end'],
+ 'from': 'prev_da_span'
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if da['value'] != value:
+ state_update['non-categorical'][-1].update({'fixed_value':da['value']})
+ ret_diff_state[domain][slot] = da['value']
+
+ break
+ if found:
+ break
+
+ if found:
+ break
+
+ # from utterance
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+ (start, end), num = pharse_in_sen(str(value), _utt)
+ if num:
+ assert value.lower() == _utt[start:end].lower() \
+ or digit2word[value].lower() == _utt[start:end].lower()
+ found = True
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': _utt[start:end].lower(),
+ # 'fixed_value': _utt[start:end].lower(),
+ 'from': 'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': start,
+ 'end': end
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if value != _utt[start:end].lower():
+ state_update['non-categorical'][-1].update({'fixed_value': _utt[start:end].lower()})
+ ret_diff_state[domain][slot] = _utt[start:end].lower()
+ found = True
+ break
+ if found:
+ break
+
+ # from utterance
+ if not value.isdigit():
+ for _turn in prev_turns[::-1]:
+ _utt = _turn['utterance']
+
+ s = difflib.SequenceMatcher(None, _utt, value)
+ matches = s.get_matching_blocks()
+
+ for i, j, n in matches:
+ possible_value = _utt[i: i+len(value)]
+
+ if i+ len(value) < len(_utt) and _utt[i+len(value)] not in [ ' ', ',', '.', '?', '!', '/'] :
+ possible_value += _utt[i+len(value):].split()[0]
+
+ if possible_value.startswith('th '):
+ possible_value = possible_value[3:]
+ i += 3
+ if i > 0 and _utt[i-1] not in [ ' ', ',', '.', '?', '!', '/']:
+ # cut first incomplete word
+ if len(possible_value.split()) > 1:
+ i += len(possible_value.split()[0]) + 1
+ possible_value = ' '.join(possible_value.split()[1:])
+
+
+ # prepend first incomplete word
+ # possible_value = _utt[:i].split()[-1] + possible_value
+ # i -= len(_utt[:i].split()[-1])
+
+
+ if fuzz.token_sort_ratio(value, possible_value) > 92 or possible_value.startswith('ashley hotel and lovell lodge') :
+ found = True
+
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ # 'value': possible_value,
+ # 'fixed_value': possible_value,
+ 'from':'prev_utt',
+ 'utt_idx': _turn['utt_idx'],
+ 'start': i,
+ 'end': i+len(possible_value)
+ })
+ if value != diff_state[domain][slot]:
+ state_update['non-categorical'][-1].update({'fixed_value': value})
+ ret_diff_state[domain][slot] = value
+ if possible_value != value:
+ state_update['non-categorical'][-1].update({'fixed_value': possible_value})
+ ret_diff_state[domain][slot] = possible_value
+ break
+ # assert _utt[i:i+len(possible_value)] == possible_value, print(_utt, _utt[i:i+len(possible_value)], possible_value)
+ # break
+ # if not possible_value == value:
+ # print(3333333333333333)
+ # print(value)
+ # print(possible_value)
+ if found:
+ break
+ if found:
+ break
+
+ if found:
+ break
+ if not found:
+ # print('3333333333333333333')
+ # print(did, tid)
+ # print(domain, slot, value)
+ # print([_t['utterance'] for _t in prev_turns])
+ # assert slot not in ['internet', 'parking', 'pricerange', 'day', 'area', 'stars']
+
+ if (domain, slot) not in slot_notfound_dict:
+ slot_notfound_dict[(domain, slot)] = 1
+ else:
+ slot_notfound_dict[(domain, slot)] += 1
+ state_update['non-categorical'].append({
+ 'domain': domain,
+ 'slot': slot,
+ 'value': diff_state[domain][slot],
+ 'fixed_value': 'not found'
+ })
+ ret_diff_state[domain][slot] = 'not found'
+ notfoundnum += 1
+ return state_update, notfoundnum, total_value, ret_diff_state
+
+
+def merge_data_annotation():
+ extract_dir = os.path.join(self_dir, 'original_data')
+ data25 = json.load(open(os.path.join(self_dir, extract_dir, 'data_meta_fixed.json')))
+ # data21_train = json.load(open(os.path.join(self_dir, extract_dir, 'train.json')))
+ # data21_val = json.load(open(os.path.join(self_dir, extract_dir, 'val.json')))
+ # data21_test = json.load(open(os.path.join(self_dir, extract_dir, 'test.json')))
+ # data21 = {}
+ # data21.update(data21_train)
+ # data21.update(data21_val)
+ # data21.update(data21_test)
+
+ # update_from_25_cnt = 0
+ # total_turn = 0
+ # for dial_id, dialog in data21.items():
+ # dial_id = dial_id + '.json'
+ # assert dial_id in data25
+ # for i, _turn in enumerate(dialog['log']):
+ # total_turn += 1
+ # if _turn['text'] == data25[dial_id]['log'][i]['text']:
+ # _turn['span_info'].extend(copy.deepcopy(data25[dial_id]['log'][i]['span_info']))
+ # # _turn['span_info'] = list(set(_turn['span_info']))
+ # # _turn['dialog_act'].update(copy.deepcopy(data25[dial_id]['log'][i]['dialog_act']))
+ # for Domain_Intent in data25[dial_id]['log'][i]['dialog_act']:
+ # if Domain_Intent in _turn['dialog_act']:
+ # _turn['dialog_act'][Domain_Intent].extend(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # else:
+ # _turn['dialog_act'][Domain_Intent] = copy.deepcopy(data25[dial_id]['log'][i]['dialog_act'][Domain_Intent])
+ # # _turn['dialog_act'][Domain_Intent] = list(set(_turn['dialog_act'][Domain_Intent]))
+ # if 'coreference' in data25[dial_id]['log'][i]:
+ # _turn['coreference'] = copy.deepcopy(data25[dial_id]['log'][i]['coreference'])
+ # update_from_25_cnt += 1
+ # else:
+ # # print('==============multiwoz21=================')
+ # # print(_turn['text'])
+ # # print('==============multiwoz25=================')
+ # # print(data25[dial_id]['log'][i]['text'])
+ # continue
+
+ # print('{}/{} turns update from multiwoz25 data'.format(update_from_25_cnt, total_turn))
+ return data25
+
+
+def preprocess(da_cat_slot_values, state_cat_slot_values):
+ all_data = []
+ binary_ont = []
+ intent_ont = {}
+ state_ont = {}
+
+ data_splits = ['train', 'val', 'test']
+ # data_splits = ['test']
+ extract_dir = os.path.join(self_dir, 'original_data')
+ num_train_dialogue = 0
+ num_train_utt = 0
+
+ num_match_error_da_span = 0
+
+ if not os.path.exists('data.zip') or not os.path.exists('ontology.json'):
+ # for data_split in data_splits:
+ data_zip_file = os.path.join(self_dir, 'original_data.zip')
+ if not os.path.exists(data_zip_file):
+ raise FileNotFoundError(data_zip_file)
+
+ logging.info('unzip multiwoz data to {}'.format(extract_dir))
+ archive = zipfile.ZipFile(data_zip_file, 'r')
+ archive.extractall(extract_dir)
+
+ data = merge_data_annotation()
+ # exit()
+ # data = json.load(open(os.path.join(self_dir, extract_dir, 'data_meta_fixed.json')))
+ train_list = open(os.path.join(self_dir, extract_dir, 'trainListFile')).read().split()
+ val_list = open(os.path.join(self_dir, extract_dir, 'valListFile')).read().split()
+ test_list = open(os.path.join(self_dir, extract_dir, 'testListFile')).read().split()
+
+ total_not_found_slot = 0
+ total_slot = 0
+ total_turn = 0
+ total_not_found_turn = 0
+ total_not_found_state = 0
+
+ slot_notfound_dict = {}
+
+ dialog_idx = 0
+ for dialog_id, dialog in tqdm(data.items()):
+
+ acc_not_found_flag = False
+
+ coref_dict = {}
+
+ data_split = None
+ for _split in data_splits:
+ if dialog_id.strip('.json') in eval(_split + '_list'):
+ data_split = _split
+ break
+ # assert data_split is not None
+ # if data_split != 'test':
+ # continue
+
+ if data_split == 'train':
+ num_train_dialogue += len(data)
+
+ dialog_idx += 1
+ # if dialog_idx > 10:
+ # break
+ converted_dialogue = {
+ 'dataset': 'multiwoz23',
+ 'data_split': data_split,
+ 'dialogue_id': 'multiwoz23_' + str(dialog_idx),
+ 'original_id': dialog_id,
+ 'domains': [d for d in dialog['goal'] if
+ len(dialog['goal'][d]) != 0 and d in multiwoz_desc and d not in ['police', 'hospital', 'bus']],
+ 'turns': [],
+ }
+
+ if data_split == 'train':
+ num_train_utt += len(dialog['log'])
+
+ prev_state = None
+ accum_fixed_state = {}
+ for turn_id, turn in enumerate(dialog['log']):
+
+ utt = turn['text'].lower()
+ # for several wrong words
+ utt = utt.replace('seeuni', 'see uni')
+
+ utt = ' '.join(utt.split())
+ das = turn['dialog_act']
+ role = 'user' if turn_id % 2 == 0 else 'system'
+ spans = turn['span_info']
+
+ da_dict = {}
+ for Domain_Act in das:
+ Domain = Domain_Act.split('-')[0]
+ if Domain.lower() not in converted_dialogue['domains'] and Domain.lower() not in ['general', 'booking']:
+ continue
+
+ Svs = das[Domain_Act]
+ for S, v in Svs:
+ v = v.lower()
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+
+ if (Domain_Act, S, v) not in da_dict:
+ da_dict[(Domain_Act, S, v)] = {}
+
+ for span in spans:
+ Domain_Act, S, v, start_word, end_word = span
+ v = v.lower()
+ if not (Domain_Act, S, v) in da_dict:
+ # logging.info('span da annotation not found in multiwoz da label')
+ # logging.info(dialog_id, turn_id)
+ # logging.info((Domain_Act, S, v))
+ # logging.info(da_dict)
+ num_match_error_da_span += 1
+ else:
+ if v.startswith('th '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+ if v.startswith('he '):
+ # print(v)
+ v = v[3:]
+ start_word += 3
+
+ if 'start_word' not in da_dict[(Domain_Act, S, v)]:
+ da_dict[(Domain_Act, S, v)]['start_word'] = []
+ da_dict[(Domain_Act, S, v)]['end_word'] = []
+
+ da_dict[(Domain_Act, S, v)]['start_word'].append(start_word)
+ da_dict[(Domain_Act, S, v)]['end_word'].append(end_word)
+
+ converted_turn = {
+ 'utt_idx': turn_id,
+ 'speaker': role,
+ 'utterance': utt,
+ 'dialogue_act': convert_da(utt, da_dict, binary_ont, intent_ont, dialog_id, turn_id, da_cat_slot_values),
+ }
+
+ # for state annotations
+ if role == 'system':
+ turn_state = turn['metadata']
+ cur_state = {}
+ for domain in turn_state:
+ if domain in ['police', 'hospital', 'bus']:
+ continue
+ if domain not in converted_dialogue['domains']:
+ continue
+ cur_state[domain] = {}
+ for subdomain in ['semi', 'book']:
+ for slot in turn_state[domain][subdomain]:
+ if slot == 'booked':
+ continue
+ if slot == 'ticket': # or (domain == 'train' and slot == 'people'):
+ # for cases where domain slot exists in REF but not in state
+ # because of check in evaluate.py
+ continue
+
+ else:
+ fixed_slot = slot
+ state_ds = domain + '-' + fixed_slot
+ if state_ds not in slot_to_type:
+ logging.info('state slot not defined in da list')
+ logging.info(state_ds)
+ if turn_state[domain][subdomain][slot] in ['', [], 'not mentioned', 'none']:
+ cur_state[domain][fixed_slot] = ""
+ else:
+ if turn_state[domain][subdomain][slot].startswith('th '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+ if turn_state[domain][subdomain][slot].startswith('he '):
+ # print('state')
+ # print(turn_state[domain][subdomain][slot])
+ turn_state[domain][subdomain][slot] = turn_state[domain][subdomain][slot][3:]
+
+ cur_state[domain][fixed_slot] = turn_state[domain][subdomain][slot]
+
+ if domain not in state_ont:
+ state_ont[domain] = []
+ if fixed_slot not in state_ont[domain]:
+ state_ont[domain].append(fixed_slot)
+
+ if domain == 'train' and 'people' not in cur_state[domain]:
+ cur_state[domain]['people'] = ''
+ # if len(converted_turn['state'][domain]) == 0:
+ # converted_turn['state'].pop(domain)
+ if len(converted_dialogue['turns']) > 0:
+ # move state from system side to user side
+ converted_dialogue['turns'][-1]['state'] = copy.deepcopy(cur_state)
+
+ # for state update annotations
+ state_update, _notfoundslot, _totalslot, ret_diff_state = get_state_update(prev_state, cur_state, converted_dialogue,
+ dialog_id, turn_id, turn['text'], coref_dict,
+ slot_notfound_dict, da_cat_slot_values)
+
+ update_state(accum_fixed_state, ret_diff_state)
+ for domain in accum_fixed_state:
+ for slot in accum_fixed_state[domain]:
+ assert isinstance(accum_fixed_state[domain][slot], str), print(accum_fixed_state[domain][slot])
+
+ if _notfoundslot == 0:
+ # for slot in state_update['categorical']:
+ # assert 'fixed_value' not in slot
+ for slot in state_update['non-categorical']:
+ if slot['value'] not in ['', 'dontcare']:
+ assert 'utt_idx' in slot
+
+ else:
+ flag = False
+ for slot in state_update['categorical']:
+ if 'fixed_value' in slot:
+ flag = True
+ break
+ for slot in state_update['non-categorical']:
+ if 'utt_idx' not in slot:
+ flag = True
+ break
+ assert flag, print(flag, state_update['non-categorical'])
+
+ total_turn += 1
+ total_slot += _totalslot
+ total_not_found_slot += _notfoundslot
+ total_not_found_turn += 1 if _notfoundslot > 0 else 0
+ if _notfoundslot > 0:
+ acc_not_found_flag = True
+ if acc_not_found_flag:
+ total_not_found_state += 1
+
+ coref_dict = {}
+ converted_dialogue['turns'][-1]['state_update'] = copy.deepcopy(state_update)
+ converted_dialogue['turns'][-1]['fixed_state'] = copy.deepcopy(accum_fixed_state)
+ if 'state' not in converted_dialogue['turns'][-1]:
+ converted_dialogue['turns'][-1]['state'] = {}
+ prev_state = copy.deepcopy(cur_state)
+
+ converted_dialogue['turns'].append(converted_turn)
+
+ if 'coreference' in turn:
+ for Domain_Act in turn['coreference']:
+ for Slot, value, coref, coref_turn, coref_pos in turn['coreference'][Domain_Act]:
+ value = value.lower()
+ coref_dict[(Domain_Act, Slot, value)] = {'turn': coref_turn, 'pos': coref_pos,
+ 'coref_value': coref,
+ 'utt': converted_dialogue['turns'][coref_turn][
+ 'utterance']}
+
+ check_spans(converted_dialogue)
+ postprocess_update_spans(converted_dialogue)
+ if converted_dialogue['turns'][-1]['speaker'] == 'system':
+ converted_dialogue['turns'].pop(-1)
+ all_data.append(converted_dialogue)
+
+ print('total_turn', total_turn)
+ print('total_not_found_turn', total_not_found_turn)
+ print('total_slot', total_slot)
+ print('total_not_found_slot', total_not_found_slot)
+ print('total_not_found_state', total_not_found_state)
+ print(slot_notfound_dict)
+ from collections import Counter
+ # print({k : dict(Counter(v)) for k, v in cat_slot_values.items()})
+ json.dump({k : dict(Counter(v)) for k, v in state_cat_slot_values.items()}, open(os.path.join(self_dir, 'cat_slot_values.json'), 'w'), indent=4)
+ cat_slot_values = {k: list(set(v)) for k, v in state_cat_slot_values.items()}
+ da_cat_slot_values = {k: list(set(v)) for k, v in da_cat_slot_values.items()}
+
+ json.dump(all_data, open('data.json', 'w'), indent=4)
+ write_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ os.remove('data.json')
+
+ new_ont = {
+ 'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': {}
+ }
+
+ for d_s in slot_to_type:
+ d, s = d_s.split('-')
+ if d not in new_ont['domains']:
+ new_ont['domains'][d] = {
+ 'description': multiwoz_desc[d]['domain'],
+ 'slots': {}
+ }
+ domain_ont = new_ont['domains'][d]
+ assert s not in domain_ont
+ domain_ont['slots'][s] = {
+ 'description': multiwoz_desc[d][s] if s in multiwoz_desc[d] else '',
+ 'is_categorical': d_s in state_cat_slot_ds,
+ 'possible_values': da_cat_slot_values[d_s] if d_s in state_cat_slot_ds else []
+ }
+ domain_ont['slots'][s]['possible_values'] = [_ for _ in domain_ont['slots'][s]['possible_values'] if _ not in ['dontcare', '']]
+
+ new_ont['state'] = {}
+ # print(state_cat_slot_value_dict)
+ print(state_ont)
+ for d in state_ont:
+ new_ont['state'][d] = {}
+ for s in state_ont[d]:
+ d_s = '-'.join([d, s])
+ new_ont['state'][d][s] = ''
+ # new_ont['state'][d][s] = {
+ # 'description': multiwoz_desc[d][s] if s in multiwoz_desc[d] else '',
+ # 'is_categorical': d_s in state_cat_slot_value_dict,
+ # 'possible_values': list(state_cat_slot_value_dict[d_s].keys()) if d_s in state_cat_slot_value_dict else []
+ # }
+ # new_ont['state'][d][s]['possible_values'] = [_ for _ in new_ont['state'][d][s]['possible_values'] if
+ # _ != 'dontcare']
+
+ new_ont['intents'] = {i: {'description': multiwoz_desc['intents'][i]} for i in intent_ont}
+ new_ont['binary_dialogue_act'] = binary_ont
+
+ slot_desc = json.load(open(os.path.join(self_dir, extract_dir, './slot_descriptions.json')))
+ for domain_slot in slot_desc:
+ _domain, _slot = domain_slot.split('-')
+ _desc = slot_desc[domain_slot][0]
+ if _slot == 'arriveby':
+ _slot = 'arriveBy'
+ elif _slot == 'leaveat':
+ _slot = 'leaveAt'
+ if 'book' in _slot:
+ _slot = _slot.replace('book ', '')
+ if not _domain in new_ont['state']:
+ # logging.info('domain {} not in state domains'.format(_domain))
+ continue
+ if _domain in new_ont['domains'] and _slot in new_ont['domains'][_domain]['slots']:
+ new_ont['domains'][_domain]['slots'][_slot]['description'] = _desc
+ if not _slot in new_ont['state'][_domain]:
+ logging.info('domain {} slot {} not in state'.format(_domain, _slot))
+ continue
+ # new_ont['state'][_domain][_slot] = ""
+ assert _domain in new_ont['domains'], print(_domain)
+ assert _slot in new_ont['domains'][_domain]['slots']
+
+ logging.info('num_match_error_da_span {}'.format(num_match_error_da_span))
+ json.dump(new_ont, open(os.path.join(self_dir, './ontology.json'), 'w'), indent=4)
+
+ else:
+ all_data = read_zipped_json(os.path.join(self_dir, './data.zip'), 'data.json')
+ new_ont = json.load(open(os.path.join(self_dir, './ontology.json'), 'r'))
+ logging.info('# dialogue: {}, # turn: {}'.format(num_train_dialogue, num_train_utt))
+ return all_data, new_ont
+
+
+def postprocess_update_spans(dialog):
+ changed_utt_idx_and_position = {}
+ for turn in dialog['turns']:
+ if turn['speaker'] != 'user':
+ continue
+ changed = False
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+
+ # assume at most one word changes for every utterance
+ if turn['utt_idx'] not in changed_utt_idx_and_position:
+ if utt_idx == turn['utt_idx'] and start-1 > -1 and turn['utterance'][start-1] not in [' ']:
+ changed_utt_idx_and_position[turn['utt_idx']] = start
+ print('=======================')
+ print(dialog['original_id'])
+ print(turn['utterance'])
+ print(json.dumps(_update, indent=2))
+ print(turn['utterance'][start: end])
+ turn['utterance'] = turn['utterance'][:start] + ' ' + turn['utterance'][start:]
+ print(turn['utterance'])
+ _update['start'] += 1
+ _update['end'] += 1
+ changed = True
+ if utt_idx not in changed_utt_idx_and_position:
+ continue
+ else:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update['fixed_value'] != 'not found' else _update['value']
+ if start >= changed_utt_idx_and_position[utt_idx]:
+ if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+ assert dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+ _update['start'] += 1
+ _update['end'] += 1
+ elif start < changed_utt_idx_and_position[utt_idx] < end:
+ if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+ assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1], value)
+ print('fix')
+ print(_update)
+ _update['end'] += 1
+ _update['fixed_value'] = turn['utterance'][_update['start']: _update['end'] + 1].strip()
+ print(_update)
+ if changed:
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+
+ if utt_idx not in changed_utt_idx_and_position:
+ continue
+ else:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+ 'fixed_value'] != 'not found' else _update['value']
+ if start >= changed_utt_idx_and_position[utt_idx]:
+ if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+ assert dialog['turns'][utt_idx]['utterance'][_update['start'] + 1: _update['end'] + 1] == value
+ _update['start'] += 1
+ _update['end'] += 1
+ elif start < changed_utt_idx_and_position[utt_idx] < end:
+ if dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']] != value:
+ print('====================fix===================')
+ print(_update)
+ assert (dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end']+1]).replace(' ', '') == value.replace(' ', ''), print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][_update['start']+1: _update['end']+1])
+ _update['end'] += 1
+ _update['fixed_value'] = dialog['turns'][utt_idx]['utterance'][_update['start']: _update['end'] + 1]
+ print(_update)
+ for turn in dialog['turns']:
+ if turn['speaker'] != 'user':
+ continue
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+ 'fixed_value'] != 'not found' else _update['value']
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+ if dialog['turns'][utt_idx]['utterance'][start] == ' ':
+ _update['start'] += 1
+ _update['fixed_value'] = value[1:]
+ value = value[1:]
+ start += 1
+ assert dialog['turns'][utt_idx]['utterance'][start: end] == value, print(json.dumps(turn, indent=4), [c for c in dialog['turns'][utt_idx]['utterance'][start: end]], [c for c in value])
+ return dialog
+
+
+def get_time_variants(time_text):
+ value_list = [time_text]
+ pattern_time = r'(\d{1,2}:\d{2})(\s)?(am|pm|AM|PM)?'
+ match_times = re.findall(pattern_time, time_text)
+ if len(match_times) < 1:
+ return []
+ match_time = match_times[0]
+
+ am_flag = match_time[2] in ['am', 'AM']
+ pm_flag = match_time[2] in ['pm', 'PM']
+ no_am_pm_flag = match_time[2] == ''
+ if am_flag:
+ # 4:00am -> 4:00
+ value_list.append(match_time[0])
+ if len(match_time[0]) == 4:
+ # 4:00 -> 04:00
+ value_list.append('0' + match_time[0])
+ if pm_flag:
+ # 4:00pm -> 16:00
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ new_hour = 12 + hour
+ value_list.append(str(new_hour)+':'+min)
+ if no_am_pm_flag:
+ hour, min = match_time[0].split(':')
+ hour = int(hour)
+ if hour > 12:
+ new_hour = hour - 12
+ value_list.append(str(new_hour) + ':' + min + 'pm')
+ value_list.append(str(new_hour) + ':' + min + ' pm')
+ value_list.append(str(new_hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(new_hour) + 'pm')
+ value_list.append(str(new_hour) + ' pm')
+ value_list.append(str(new_hour))
+ else:
+ value_list.append(str(hour) + ':' + min + 'am')
+ value_list.append(str(hour) + ':' + min + ' am')
+ value_list.append(str(hour) + ':' + min)
+ if min == '00':
+ value_list.append(str(hour) + 'am')
+ value_list.append(str(hour) + ' am')
+ value_list.append(str(hour))
+ if len(match_time[0]) == 5 and match_time[0][0] == '0':
+ value_list.append(match_time[0][1:])
+ value_list.append(''.join(match_time[0].split(':')))
+
+ return value_list
+
+
+def get_genitive_variants(value):
+ ret_list = []
+ value_genitive_format = r"(?=\w)s(?=\s)"
+ value_pattern = re.compile(value_genitive_format)
+
+ span_genitive_value = re.sub(value_pattern, " 's", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ span_genitive_value = re.sub(value_pattern, "'s", value)
+ if span_genitive_value != value:
+ ret_list.append(span_genitive_value)
+ # if len(ret_list) > 0:
+ # print('=============================')
+ # print(value)
+ # print(re.findall(value_pattern, value))
+ # print(ret_list)
+ return ret_list
+
+
+def check_spans(dialog):
+ for turn in dialog['turns']:
+ if turn['speaker'] != 'user':
+ continue
+ for _update in turn['state_update']['non-categorical']:
+ if 'utt_idx' in _update:
+ value = _update['fixed_value'] if 'fixed_value' in _update and _update[
+ 'fixed_value'] != 'not found' else _update['value']
+ utt_idx = _update['utt_idx']
+ start = _update['start']
+ end = _update['end']
+ assert dialog['turns'][utt_idx]['utterance'][start:end] == value, print(dialog['turns'][utt_idx]['utterance'], dialog['turns'][utt_idx]['utterance'][start:end])
+
+
+
+def get_bb_variants(value):
+ ret_list = []
+ if 'bed and breakfast' in value:
+ ret_list.append(value.replace('bed and breakfast', 'b & b'))
+ return ret_list
+
+if __name__ == '__main__':
+ preprocess(da_cat_slot_values, state_cat_slot_values)
\ No newline at end of file
diff --git a/data/unified_datasets/schema/README.md b/data/unified_datasets/schema/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..32483a533a248fda59a92dfeceab436d741659de
--- /dev/null
+++ b/data/unified_datasets/schema/README.md
@@ -0,0 +1,689 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, belief state, character-level span for non-categorical slots.
+- Unseen domains and slots in the test set to quantify the performance in zero-shot or few shot settings.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 16142 | 313822 | 19.44 | 10.02 | 16 |
+| val | 2482 | 46244 | 18.63 | 9.94 | 16 |
+| test | 4201 | 80393 | 19.14 | 10.7 | 18 |
+
+## Main changes
+
+1. download the original data as `original_data.zip`
+
+2. run `python preprocess` to unzip `original_data.zip` and get processed `data.zip` & `ontology.json`.
+
+Main changes:
+
+- extract intent from domains.
+- ~~numerical slot => non-categorical, use string match to get the span.~~
+- add binary_dialogue_act for those binary intents such as 'goodbye', 'request'.
+- add **count** non-categorical, numerical slot for each domain, but not appear in belief state.
+- sys state are updated by previous user frame['state'].
+- calculate the state update according to prev state and slot spans in current turn slot_vals and all previous dialogue acts. 99.6% non-categorical state update have spans while the rest of them are like "Could you help me search for songs from **two years back** too?"
+- values in possible values, dialogue act, state, and state_update are in **lowercase**.
+
+Notice:
+
+- for categorical slot, value maybe **dontcare**, which is not presented in **possible_values**.
+
+## Original data
+
+The Schema-Guided Dialogue (SGD) dataset consists of over 20k annotated
+multi-domain, task-oriented conversations between a human and a virtual
+assistant. These conversations involve interactions with services and APIs
+spanning 20 domains, ranging from banks and events to media, calendar, travel,
+and weather. For most of these domains, the dataset contains multiple different
+APIs, many of which have overlapping functionalities but different interfaces,
+which reflects common real-world scenarios. The wide range of available
+annotations can be used for intent prediction, slot filling, dialogue state
+tracking, policy imitation learning, language generation, user simulation
+learning, among other tasks in large-scale virtual assistants. Besides these,
+the dataset has unseen domains and services in the evaluation set to quantify
+the performance in zero-shot or few shot settings.
+
+[[paper]](https://arxiv.org/abs/1909.05855) [[download link]](https://github.com/google-research-datasets/dstc8-schema-guided-dialogue)
+
+### Scheme Representation
+
+A service or API is essentially a set of functions (called intents), each taking
+a set of parameters (called slots). A schema is a normalized representation of
+the interface exposed by a service/API. In addition, the schema also includes
+natural language description of the included functions and their parameters to
+outline the semantics of each element. The schemas have been manually generated
+by the dataset creators. The schema for a service contains the following fields:
+
+* **service_name** - A unique name for the service.
+* **description** - A natural language description of the tasks supported by
+ the service.
+* **slots** - A list of slots/attributes corresponding to the entities present
+ in the service. Each slot contains the following fields:
+ * **name** - The name of the slot.
+ * **description** - A natural language description of the slot.
+ * **is_categorical** - A boolean value. If it is true, the slot has a
+ fixed set of possible values.
+ * **possible_values** - List of possible values the slot can take. If the
+ slot is a categorical slot, it is a complete list of all the possible
+ values. If the slot is a non categorical slot, it is either an empty
+ list or a small sample of all the values taken by the slot.
+* **intents** - The list of intents/tasks supported by the service. Each
+ method contains the following fields:
+ * **name** - The name of the intent.
+ * **description** - A natural language description of the intent.
+ * **is_transactional** - A boolean value. If true, indicates that the
+ underlying API call is transactional (e.g, a booking or a purchase), as
+ opposed to a search call.
+ * **required_slots** - A list of slot names whose values must be provided
+ before making a call to the service.
+ * **optional_slots** - A dictionary mapping slot names to the default
+ value taken by the slot. These slots may be optionally specified by the
+ user and the user may override the default value. An empty default value
+ allows that slot to take any value by default, but the user may override
+ it.
+ * **result_slots** - A list of slot names which are present in the results
+ returned by a call to the service or API.
+
+### Dialogue Representation
+
+The dialogue is represented as a list of turns, where each turn contains either
+a user or a system utterance. The annotations for a turn are grouped into
+frames, where each frame corresponds to a single service. Each turn in the
+single domain dataset contains exactly one frame. In multi-domain datasets, some
+turns may have multiple frames.
+
+Each dialogue is represented as a json object with the following fields:
+
+* **dialogue_id** - A unique identifier for a dialogue.
+* **services** - A list of services present in the dialogue.
+* **turns** - A list of annotated system or user utterances.
+
+Each turn consists of the following fields:
+
+* **speaker** - The speaker for the turn. Possible values are "USER" or
+ "SYSTEM".
+* **utterance** - A string containing the natural language utterance.
+* **frames** - A list of frames, each frame containing annotations for a
+ single service.
+
+Each frame consists of the fields listed below. The fields marked with * will
+be excluded from all user turns in the test data released to the participants.
+
+* **service** - The name of the service corresponding to the frame. The slots
+ and intents used in the following fields are taken from the schema of this
+ service.
+* **slots** - A list of slot spans in the utterance, only provided for
+ non-categorical slots. Each slot span contains the following fields:
+ * **slot** - The name of the slot.
+ * **start** - The index of the starting character in the utterance
+ corresponding to the slot value.
+ * **exclusive_end** - The index of the character just after the last
+ character corresponding to the slot value in the utterance. In python,
+ `utterance[start:exclusive_end]` gives the slot value.
+* **actions** - A list of actions corresponding to the system. Each action has
+ the following fields:
+ * **act** - The type of action. The list of all possible system acts is
+ given below.
+ * **slot** (optional) - A slot argument for some of the actions.
+ * **values** (optional) - A list of values assigned to the slot. If the
+ values list is non-empty, then the slot must be present.
+ * **canonical_values** (optional) - The values in their canonicalized form
+ as used by the service. It is a list of strings of the same length as
+ values.
+* **service_call** (system turns only, optional) - The request sent to the
+ service. It consists of the following fields:
+ * **method** - The name of the intent or function of the service or API
+ being executed.
+ * **parameters** - A dictionary mapping slot name (all required slots and
+ possibly some optional slots) to a value in its canonicalized form.
+* **service_results** (system turns only, optional) - A list of entities
+ containing the results obtained from the service. It is only available for
+ turns in which a service call is made. Each entity is represented as a
+ dictionary mapping a slot name to a string containing its canonical value.
+* **state** (user turns only) - The dialogue state corresponding to the
+ service. It consists of the following fields:
+ * **active_intent** - The intent corresponding to the service of the frame
+ which is currently being fulfilled by the system. It takes the value
+ "NONE" if none of the intents are active.
+ * **requested_slots** - A list of slots requested by the user in the
+ current turn.
+ * **slot_values** - A dictionary mapping slot name to a list of strings.
+ For categorical slots, this list contains a single value assigned to the
+ slot. For non-categorical slots, all the values in this list are spoken
+ variations of each other and are equivalent (e.g, "6 pm", "six in the
+ evening", "evening at 6" etc.).
+
+List of possible system acts:
+
+* **INFORM** - Inform the value for a slot to the user. The slot and values
+ fields in the corresponding action are always non-empty.
+* **REQUEST** - Request the value of a slot from the user. The corresponding
+ action always contains a slot, but values are optional. When values are
+ present, they are used as examples for the user e.g, "Would you like to eat
+ indian or chinese food or something else?"
+* **CONFIRM** - Confirm the value of a slot before making a transactional
+ service call.
+* **OFFER** - Offer a certain value for a slot to the user. The corresponding
+ action always contains a slot and a list of values for that slot offered to
+ the user.
+* **NOTIFY_SUCCESS** - Inform the user that their request was successful. Slot
+ and values are always empty in the corresponding action.
+* **NOTIFY_FAILURE** - Inform the user that their request failed. Slot and
+ values are always empty in the corresponding action.
+* **INFORM_COUNT** - Inform the number of items found that satisfy the user's
+ request. The corresponding action always has "count" as the slot, and a
+ single element in values for the number of results obtained by the system.
+* **OFFER_INTENT** - Offer a new intent to the user. Eg, "Would you like to
+ reserve a table?". The corresponding action always has "intent" as the slot,
+ and a single value containing the intent being offered. The offered intent
+ belongs to the service corresponding to the frame.
+* **REQ_MORE** - Asking the user if they need anything else. Slot and values
+ are always empty in the corresponding action.
+* **GOODBYE** - End the dialogue. Slot and values are always empty in the
+ corresponding action.
+
+List of possible user acts:
+
+* **INFORM_INTENT** - Express the desire to perform a certain task to the
+ system. The action always has "intent" as the slot and a single value
+ containing the intent being informed.
+* **NEGATE_INTENT** - Negate the intent which has been offered by the system.
+* **AFFIRM_INTENT** - Agree to the intent which has been offered by the
+ system.
+* **INFORM** - Inform the value of a slot to the system. The slot and values
+ fields in the corresponding action are always non-empty.
+* **REQUEST** - Request the value of a slot from the system. The corresponding
+ action always contains a slot parameter. It may optionally contain a value,
+ in which case, the user asks the system if the slot has the specified value.
+* **AFFIRM** - Agree to the system's proposition. Slot and values are always
+ empty.
+* **NEGATE** - Deny the system's proposal. Slot and values are always empty.
+* **SELECT** - Select a result being offered by the system. The corresponding
+ action may either contain no parameters, in which case all the values
+ proposed by the system are being accepted, or it may contain a slot and
+ value parameters, in which case the specified slot and value are being
+ accepted.
+* **REQUEST_ALTS** - Ask for more results besides the ones offered by the
+ system. Slot and values are always empty.
+* **THANK_YOU** - Thank the system. Slot and values are always empty.
+* **GOODBYE** - End the dialogue. Slot and values are always empty.
+
+### Dataset Statistics
+
+The dataset consists of two kinds of dialogues.
+
+| Type of Dialogue | Train files | Dev files | Test Files |
+| ---------------- | :------------------------------------------: | :------------------------------------------: | :------------------------------------------: |
+| Single Domain | `dialogues_001.json` to `dialogues_043.json` | `dialogues_001.json` to `dialogues_007.json` | `dialogues_001.json` to `dialogues_011.json` |
+| Multi Domain | `dialogues_044.json` to `dialogues_127.json` | `dialogues_008.json` to `dialogues_020.json` | `dialogues_012.json` to `dialogues_034.json` |
+
+The single domain dialogues involve interactions with a single service, possibly
+over multiple intents. The multi-domain dialogues have interactions involving
+intents belonging to two or more different services. The multi-domain dialogues
+also involve transfer of dialogue state values from one service to the other
+wherever such a transfer is deemed natural. Eg, if a user finds a restaurant and
+searches for a movie next, the dialogue state for movie service is already
+initialized with the location from the dialogue state for restaurant service.
+
+The overall statistics of the train and dev sets are given below. The term
+*informable slots* refers to the slots over which the user can specify a
+constraint. For example, slots like *phone_number* are not informable.
+
+<table>
+ <tr>
+ <th rowspan="2"></th>
+ <th colspan="3">Train</th><th colspan="3">Dev</th><th colspan="3">Test</th>
+ </tr>
+ <tr>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ </tr>
+ <tr>
+ <td>No. of dialogues</td>
+ <td align="center">5,403</td>
+ <td align="center">10,739</td>
+ <td align="center">16,142</td>
+ <td align="center">836</td>
+ <td align="center">1,646</td>
+ <td align="center">2,482</td>
+ <td align="center">1,331</td>
+ <td align="center">2,870</td>
+ <td align="center">4,201</td>
+ </tr>
+ <tr>
+ <td>No. of turns</td>
+ <td align="center">82,588</td>
+ <td align="center">247,376</td>
+ <td align="center">329,964</td>
+ <td align="center">11,928</td>
+ <td align="center">36,798</td>
+ <td align="center">48,726</td>
+ <td align="center">16,850</td>
+ <td align="center">67,744</td>
+ <td align="center">84,594</td>
+ </tr>
+ <tr>
+ <td>No. of tokens (lower-cased)</td>
+ <td align="center">807,562</td>
+ <td align="center">2,409,857</td>
+ <td align="center">3,217,419</td>
+ <td align="center">117,492</td>
+ <td align="center">353,381</td>
+ <td align="center">470,873</td>
+ <td align="center">166,329</td>
+ <td align="center">713,731</td>
+ <td align="center">880,060</td>
+ </tr>
+ <tr>
+ <td>Average turns per dialogue</td>
+ <td align="center">15.286</td>
+ <td align="center">23.035</td>
+ <td align="center">20.441</td>
+ <td align="center">14.268</td>
+ <td align="center">22.356</td>
+ <td align="center">19.632</td>
+ <td align="center">12.660</td>
+ <td align="center">23.604</td>
+ <td align="center">20.137</td>
+ </tr>
+ <tr>
+ <td>Average tokens per turn</td>
+ <td align="center">9.778</td>
+ <td align="center">9.742</td>
+ <td align="center">9.751</td>
+ <td align="center">9.850</td>
+ <td align="center">9.603</td>
+ <td align="center">9.664</td>
+ <td align="center">9.871</td>
+ <td align="center">10.536</td>
+ <td align="center">10.403</td>
+ </tr>
+ <tr>
+ <td>Total unique tokens (lower-cased)</td>
+ <td align="center">16,350</td>
+ <td align="center">25,459</td>
+ <td align="center">30,349</td>
+ <td align="center">6,803</td>
+ <td align="center">10,533</td>
+ <td align="center">12,719</td>
+ <td align="center">7,213</td>
+ <td align="center">14,888</td>
+ <td align="center">16,382</td>
+ </tr>
+ <tr>
+ <td>Total no. of slots</td>
+ <td align="center">201</td>
+ <td align="center">214</td>
+ <td align="center">214</td>
+ <td align="center">134</td>
+ <td align="center">132</td>
+ <td align="center">136</td>
+ <td align="center">157</td>
+ <td align="center">158</td>
+ <td align="center">159</td>
+ </tr>
+ <tr>
+ <td>Total no. of informable slots</td>
+ <td align="center">138</td>
+ <td align="center">144</td>
+ <td align="center">144</td>
+ <td align="center">89</td>
+ <td align="center">87</td>
+ <td align="center">89</td>
+ <td align="center">109</td>
+ <td align="center">110</td>
+ <td align="center">111</td>
+ </tr>
+ <tr>
+ <td>Total unique slot values (lower-cased)</td>
+ <td align="center">7,070</td>
+ <td align="center">11,635</td>
+ <td align="center">14,139</td>
+ <td align="center">2,418</td>
+ <td align="center">4,182</td>
+ <td align="center">5,101</td>
+ <td align="center">2,492</td>
+ <td align="center">5,847</td>
+ <td align="center">6,533</td>
+ </tr>
+ <tr>
+ <td>Total unique informable slot values (lower-cased)</td>
+ <td align="center">3,742</td>
+ <td align="center">6,348</td>
+ <td align="center">7,661</td>
+ <td align="center">1,137</td>
+ <td align="center">2,118</td>
+ <td align="center">2,524</td>
+ <td align="center">1,387</td>
+ <td align="center">3,323</td>
+ <td align="center">3,727</td>
+ </tr>
+ <tr>
+ <td>Total domains</td>
+ <td align="center">14</td>
+ <td align="center">16</td>
+ <td align="center">16</td>
+ <td align="center">16</td>
+ <td align="center">15</td>
+ <td align="center">16</td>
+ <td align="center">17</td>
+ <td align="center">18</td>
+ <td align="center">18</td>
+ </tr>
+ <tr>
+ <td>Total services</td>
+ <td align="center">24</td>
+ <td align="center">26</td>
+ <td align="center">26</td>
+ <td align="center">17</td>
+ <td align="center">16</td>
+ <td align="center">17</td>
+ <td align="center">20</td>
+ <td align="center">21</td>
+ <td align="center">21</td>
+ </tr>
+ <tr>
+ <td>Total intents</td>
+ <td align="center">35</td>
+ <td align="center">37</td>
+ <td align="center">37</td>
+ <td align="center">28</td>
+ <td align="center">26</td>
+ <td align="center">28</td>
+ <td align="center">33</td>
+ <td align="center">34</td>
+ <td align="center">35</td>
+ </tr>
+</table>
+
+
+The following table shows how the dialogues and services are distributed among
+different domains for the train and dev sets. In this table, each multi-domain
+dialogue contirbutes to the count of every service present in the dialogue.
+Please note that a few domains like *Travel* and *Weather* are only present in
+the dev set. This is to test the generalization of models on unseen domains. The
+test set will similarly have some unseen domains which are neither present in
+the training nor in the dev set. Also, the number in parenthesis represents the
+number of unique services belonging to the corresponding domain.
+
+* In the first column, it indicates the number of unique services for the
+ domain in Train, Dev and Test datasets combined.
+* In the fourth column, it indicates the number of such unique services in the
+ Train dataset only.
+* In the seventh column, it indicates the number of such unique services in
+ the Dev dataset only.
+* In the last column, it indicates the number of such unique services in the
+ Test dataset only.
+
+<table>
+ <tr>
+ <th rowspan="2"></th>
+ <th colspan="3"># Dialogues <br> Train</th>
+ <th colspan="3"># Dialogues <br> Dev</th>
+ <th colspan="3"># Dialogues <br> Test</th>
+ </tr>
+ <tr>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ <td>Single-domain</td>
+ <td>Multi-domain</td>
+ <td>Combined</td>
+ </tr>
+ <tr>
+ <td>Alarm (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">37</td>
+ <td align="center">NA</td>
+ <td align="center">37 (1)</td>
+ <td align="center">47</td>
+ <td align="center">240</td>
+ <td align="center">287 (1)</td>
+ </tr>
+ <tr>
+ <td>Banks (2)</td>
+ <td align="center">207</td>
+ <td align="center">520</td>
+ <td align="center">727 (1)</td>
+ <td align="center">42</td>
+ <td align="center">252</td>
+ <td align="center">294 (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ </tr>
+ <tr>
+ <td>Buses (3)</td>
+ <td align="center">310</td>
+ <td align="center">1,970</td>
+ <td align="center">2,280 (2)</td>
+ <td align="center">44</td>
+ <td align="center">285</td>
+ <td align="center">329 (1)</td>
+ <td align="center">88</td>
+ <td align="center">438</td>
+ <td align="center">526 (1)</td>
+ </tr>
+ <tr>
+ <td>Calendar (1)</td>
+ <td align="center">169</td>
+ <td align="center">1,433</td>
+ <td align="center">1,602 (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ </tr>
+ <tr>
+ <td>Events (3)</td>
+ <td align="center">788</td>
+ <td align="center">2,721</td>
+ <td align="center">3,509 (1)</td>
+ <td align="center">73</td>
+ <td align="center">345</td>
+ <td align="center">418 (1)</td>
+ <td align="center">76</td>
+ <td align="center">516</td>
+ <td align="center">592 (1)</td>
+ </tr>
+ <tr>
+ <td>Flights (4)</td>
+ <td align="center">985</td>
+ <td align="center">1,762</td>
+ <td align="center">2,747 (2)</td>
+ <td align="center">94</td>
+ <td align="center">297</td>
+ <td align="center">391 (1)</td>
+ <td align="center">87</td>
+ <td align="center">419</td>
+ <td align="center">506 (1)</td>
+ </tr>
+ <tr>
+ <td>Homes (2)</td>
+ <td align="center">268</td>
+ <td align="center">579</td>
+ <td align="center">847 (1)</td>
+ <td align="center">81</td>
+ <td align="center">99</td>
+ <td align="center">180 (1)</td>
+ <td align="center">89</td>
+ <td align="center">157</td>
+ <td align="center">246 (1)</td>
+ </tr>
+ <tr>
+ <td>Hotels (4)</td>
+ <td align="center">457</td>
+ <td align="center">2,896</td>
+ <td align="center">3,353 (3)</td>
+ <td align="center">56</td>
+ <td align="center">521</td>
+ <td align="center">577 (2)</td>
+ <td align="center">177</td>
+ <td align="center">885</td>
+ <td align="center">1062 (2)</td>
+ </tr>
+ <tr>
+ <td>Media (3)</td>
+ <td align="center">281</td>
+ <td align="center">832</td>
+ <td align="center">1,113 (1)</td>
+ <td align="center">46</td>
+ <td align="center">133</td>
+ <td align="center">179 (1)</td>
+ <td align="center">80</td>
+ <td align="center">284</td>
+ <td align="center">364 (1)</td>
+ </tr>
+ <tr>
+ <td>Messaging (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">298</td>
+ <td align="center">298 (1)</td>
+ </tr>
+ <tr>
+ <td>Movies (2)</td>
+ <td align="center">292</td>
+ <td align="center">1,325</td>
+ <td align="center">1,617 (1)</td>
+ <td align="center">47</td>
+ <td align="center">94</td>
+ <td align="center">141 (1)</td>
+ <td align="center">132</td>
+ <td align="center">449</td>
+ <td align="center">581</td>
+ </tr>
+ <tr>
+ <td>Music (3)</td>
+ <td align="center">394</td>
+ <td align="center">896</td>
+ <td align="center">1,290 (2)</td>
+ <td align="center">35</td>
+ <td align="center">161</td>
+ <td align="center">196 (1)</td>
+ <td align="center">25</td>
+ <td align="center">322</td>
+ <td align="center">347 (2)</td>
+ </tr>
+ <tr>
+ <td>Payment (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">36</td>
+ <td align="center">186</td>
+ <td align="center">222 (1)</td>
+ </tr>
+ <tr>
+ <td>RentalCars (3)</td>
+ <td align="center">215</td>
+ <td align="center">1,370</td>
+ <td align="center">1,585 (2)</td>
+ <td align="center">39</td>
+ <td align="center">342</td>
+ <td align="center">381 (1)</td>
+ <td align="center">64</td>
+ <td align="center">480</td>
+ <td align="center">544 (1)</td>
+ </tr>
+ <tr>
+ <td>Restaurants (2)</td>
+ <td align="center">367</td>
+ <td align="center">2052</td>
+ <td align="center">2,419 (1)</td>
+ <td align="center">73</td>
+ <td align="center">263</td>
+ <td align="center">336 (1)</td>
+ <td align="center">73</td>
+ <td align="center">390</td>
+ <td align="center">463 (1)</td>
+ </tr>
+ <tr>
+ <td>RideSharing (2)</td>
+ <td align="center">119</td>
+ <td align="center">1,584</td>
+ <td align="center">1,703 (2)</td>
+ <td align="center">45</td>
+ <td align="center">225</td>
+ <td align="center">270 (1)</td>
+ <td align="center">34</td>
+ <td align="center">216</td>
+ <td align="center">250 (1)</td>
+ </tr>
+ <tr>
+ <td>Services (4)</td>
+ <td align="center">551</td>
+ <td align="center">1,338</td>
+ <td align="center">1,889 (3)</td>
+ <td align="center">44</td>
+ <td align="center">157</td>
+ <td align="center">201 (1)</td>
+ <td align="center">167</td>
+ <td align="center">489</td>
+ <td align="center">656 (2)</td>
+ </tr>
+ <tr>
+ <td>Trains (1)</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">NA</td>
+ <td align="center">84</td>
+ <td align="center">266</td>
+ <td align="center">350 (1)</td>
+ </tr>
+ <tr>
+ <td>Travel (1)</td>
+ <td align="center">NA</td>
+ <td align="center">1,871</td>
+ <td align="center">1,871 (1)</td>
+ <td align="center">45</td>
+ <td align="center">238</td>
+ <td align="center">283 (1)</td>
+ <td align="center">24</td>
+ <td align="center">630</td>
+ <td align="center">654 (1)</td>
+ </tr>
+ <tr>
+ <td>Weather (1)</td>
+ <td align="center">NA</td>
+ <td align="center">951</td>
+ <td align="center">951 (1)</td>
+ <td align="center">35</td>
+ <td align="center">322</td>
+ <td align="center">357 (1)</td>
+ <td align="center">48</td>
+ <td align="center">427</td>
+ <td align="center">475 (1)</td>
+ </tr>
+</table>
+
diff --git a/data/unified_datasets/schema/data.zip b/data/unified_datasets/schema/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..d9cecdf51bc1793bed81920c5953bbb7bc001699
Binary files /dev/null and b/data/unified_datasets/schema/data.zip differ
diff --git a/data/unified_datasets/schema/ontology.json b/data/unified_datasets/schema/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..56b66390614089b661d772d6fb70f75f9ce10ad1
--- /dev/null
+++ b/data/unified_datasets/schema/ontology.json
@@ -0,0 +1,7163 @@
+{
+ "domains": {
+ "bank_1": {
+ "description": "Manage bank accounts and transfer money",
+ "slots": {
+ "account_type": {
+ "description": "The account type of the user",
+ "is_categorical": true,
+ "possible_values": [
+ "checking",
+ "savings"
+ ]
+ },
+ "recipient_account_type": {
+ "description": "The account type of the recipient whom the user is transfering money to",
+ "is_categorical": true,
+ "possible_values": [
+ "checking",
+ "savings"
+ ]
+ },
+ "balance": {
+ "description": "The amount of money held in a bank account at a given time",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "amount": {
+ "description": "The amount of money to transfer",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "recipient_account_name": {
+ "description": "The account name of the recipient who is to receive the transfered money",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "bus_1": {
+ "description": "Book bus journeys from the biggest bus network in the country",
+ "slots": {
+ "from_location": {
+ "description": "City where bus is leaving from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to_location": {
+ "description": "City where bus is going to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "from_station": {
+ "description": "Station where bus is leaving from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to_station": {
+ "description": "Station where bus is going to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaving_date": {
+ "description": "Date of bus leaving for journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "leaving_time": {
+ "description": "Time of bus leaving for journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "fare": {
+ "description": "Fare per ticket for journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "travelers": {
+ "description": "Number of travelers for journey",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "transfers": {
+ "description": "Number of transfers in journey",
+ "is_categorical": true,
+ "possible_values": [
+ "0",
+ "1"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "bus_2": {
+ "description": "Find a bus to take you to the city you want",
+ "slots": {
+ "origin": {
+ "description": "Origin city for journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination": {
+ "description": "Destination city for journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "origin_station_name": {
+ "description": "Name of the bus terminus at the origin",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_station_name": {
+ "description": "Name of the bus terminus at the destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "Date of bus departure",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Price per ticket of the itinerary",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_time": {
+ "description": "Time of bus departure",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "group_size": {
+ "description": "Size of group for the booking",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "fare_type": {
+ "description": "Type of fare for the booking",
+ "is_categorical": true,
+ "possible_values": [
+ "economy",
+ "economy extra",
+ "flexible"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "calendar_1": {
+ "description": "Calendar service to manage personal events and reservations",
+ "slots": {
+ "event_date": {
+ "description": "Date of event or for checking availability",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_time": {
+ "description": "Start time of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_location": {
+ "description": "Location of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_name": {
+ "description": "Title of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "available_start_time": {
+ "description": "Starting time of available time slot",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "available_end_time": {
+ "description": "Ending time of available time slot",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "event_1": {
+ "description": "The comprehensive portal to find and reserve seats at events near you",
+ "slots": {
+ "category": {
+ "description": "Type of event",
+ "is_categorical": true,
+ "possible_values": [
+ "music",
+ "sports"
+ ]
+ },
+ "subcategory": {
+ "description": "Subcategory of event, either a music genre or sport name",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_name": {
+ "description": "Name of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Date of occurrence of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "Time when the event is scheduled to start",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_seats": {
+ "description": "Number of seats to find event tickets for",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6",
+ "7",
+ "8",
+ "9"
+ ]
+ },
+ "city_of_event": {
+ "description": "City where event is happening",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_location": {
+ "description": "Name of event location",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address_of_location": {
+ "description": "Street address of event location",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "event_2": {
+ "description": "Get tickets for the coolest concerts and sports in your area",
+ "slots": {
+ "event_type": {
+ "description": "Type of event",
+ "is_categorical": true,
+ "possible_values": [
+ "music",
+ "sports"
+ ]
+ },
+ "category": {
+ "description": "The sport or music subcategory",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "event_name": {
+ "description": "Name of match or artist for event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Date of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "Starting time for event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_tickets": {
+ "description": "Number of tickets to buy for event",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6",
+ "7",
+ "8",
+ "9"
+ ]
+ },
+ "city": {
+ "description": "City where the event is taking place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "venue": {
+ "description": "Exact venue of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "venue_address": {
+ "description": "Address of event venue",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "flight_1": {
+ "description": "Find your next flight",
+ "slots": {
+ "passengers": {
+ "description": "Number of passengers in the booking",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "seating_class": {
+ "description": "Seating class for the booking",
+ "is_categorical": true,
+ "possible_values": [
+ "economy",
+ "premium economy",
+ "business",
+ "first class"
+ ]
+ },
+ "origin_city": {
+ "description": "Starting city for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_city": {
+ "description": "Ending city for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "origin_airport": {
+ "description": "Airport at the starting city for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_airport": {
+ "description": "Airport at the ending city for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "Start date for the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "return_date": {
+ "description": "Date of the return flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_stops": {
+ "description": "Number of layovers in the flight",
+ "is_categorical": true,
+ "possible_values": [
+ "0",
+ "1"
+ ]
+ },
+ "outbound_departure_time": {
+ "description": "Departure time for the outbound leg flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "outbound_arrival_time": {
+ "description": "Arrival time for the outbound leg flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_arrival_time": {
+ "description": "Arrival time for the return leg flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_departure_time": {
+ "description": "Departure time for the return leg flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Price of the booking per passenger",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "refundable": {
+ "description": "Whether the booking is refundable or not",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "airlines": {
+ "description": "Name of airline",
+ "is_categorical": true,
+ "possible_values": [
+ "united airlines",
+ "american airlines",
+ "delta airlines",
+ "southwest airlines",
+ "alaska airlines",
+ "british airways",
+ "air canada",
+ "air france"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "flight_2": {
+ "description": "Search for cheap flights across multiple providers",
+ "slots": {
+ "passengers": {
+ "description": "Number of passengers to book tickets for",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "seating_class": {
+ "description": "Seating class for the flight tickets",
+ "is_categorical": true,
+ "possible_values": [
+ "economy",
+ "premium economy",
+ "business",
+ "first class"
+ ]
+ },
+ "origin": {
+ "description": "City of origin for the flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination": {
+ "description": "City of destination for the flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "origin_airport": {
+ "description": "Airport where the flight is departing from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_airport": {
+ "description": "Airport where the flight is arriving to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "Date of departure flight on the ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "return_date": {
+ "description": "Date of return flight on the ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_stops": {
+ "description": "Number of stopovers made by the flight",
+ "is_categorical": true,
+ "possible_values": [
+ "0",
+ "1"
+ ]
+ },
+ "outbound_departure_time": {
+ "description": "Departure time of the origin-destination flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "outbound_arrival_time": {
+ "description": "Arrival time of the origin-destination flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_arrival_time": {
+ "description": "Arrival time of the destination-origin flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_departure_time": {
+ "description": "Departure time of the destination-origin flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "fare": {
+ "description": "Ticket fare for each passenger",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "is_redeye": {
+ "description": "Boolean flag indicating whether the flight is a red-eye flight",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "airlines": {
+ "description": "Name of the airline to book with",
+ "is_categorical": true,
+ "possible_values": [
+ "united airlines",
+ "american airlines",
+ "delta airlines",
+ "southwest airlines",
+ "alaska airlines",
+ "british airways",
+ "air canada",
+ "air france"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "home_1": {
+ "description": "A widely used service for finding apartments and scheduling visits",
+ "slots": {
+ "area": {
+ "description": "City where the apartment is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Address of the apartment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "property_name": {
+ "description": "Name of the apartment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the apartment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "furnished": {
+ "description": "Boolean flag indicating if the property is furnished",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "pets_allowed": {
+ "description": "Boolean flag indicating if pets are allowed",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "rent": {
+ "description": "Rent per month of the apartment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "visit_date": {
+ "description": "Date for the visit to the apartment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_beds": {
+ "description": "Number of bed rooms",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "number_of_baths": {
+ "description": "Number of baths in the apartment",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel_1": {
+ "description": "A popular service for searching and reserving rooms in hotels",
+ "slots": {
+ "destination": {
+ "description": "Location of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_rooms": {
+ "description": "Number of rooms in the reservation",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3"
+ ]
+ },
+ "check_in_date": {
+ "description": "Start date for the reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_days": {
+ "description": "Number of days in the reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "star_rating": {
+ "description": "Star rating of the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "hotel_name": {
+ "description": "Name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Address of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price_per_night": {
+ "description": "Price per night for the reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "has_wifi": {
+ "description": "Boolean flag indicating if the hotel has wifi",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel_2": {
+ "description": "A popular service for searching and booking houses for short term stay",
+ "slots": {
+ "where_to": {
+ "description": "Location of the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_adults": {
+ "description": "Number of people for the reservation",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "check_in_date": {
+ "description": "Start date for the reservation or to find the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "check_out_date": {
+ "description": "End date for the reservation or to find the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "rating": {
+ "description": "Review rating of the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Address of the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "total_price": {
+ "description": "Price per night of the house",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "has_laundry_service": {
+ "description": "Boolean flag indicating if the house has laundry service",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel_3": {
+ "description": "A leading provider for searching and booking hotel rooms",
+ "slots": {
+ "location": {
+ "description": "Location of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_rooms": {
+ "description": "Number of rooms to book",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3"
+ ]
+ },
+ "check_in_date": {
+ "description": "Start date for the hotel reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "check_out_date": {
+ "description": "End date for the hotel reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "average_rating": {
+ "description": "Average review rating for the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "hotel_name": {
+ "description": "Name of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Address of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the hotel",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Total price for the stay",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pets_welcome": {
+ "description": "Boolean flag indicating if pets are allowed in the hotel",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "media_1": {
+ "description": "A leading provider of movies for searching and watching on-demand",
+ "slots": {
+ "title": {
+ "description": "Title of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Genre of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "subtitles": {
+ "description": "Boolean flag indicating if subtitles are desired for this movie",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "directed_by": {
+ "description": "Name of the director of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "movie_1": {
+ "description": "A go-to provider for finding movies, searching for show times and booking tickets",
+ "slots": {
+ "price": {
+ "description": "Price per ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_tickets": {
+ "description": "Number of the movie tickets to be purchased",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6",
+ "7",
+ "8",
+ "9"
+ ]
+ },
+ "show_type": {
+ "description": "Type of show",
+ "is_categorical": true,
+ "possible_values": [
+ "regular",
+ "3d",
+ "imax"
+ ]
+ },
+ "theater_name": {
+ "description": "Name of the theatre",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "show_time": {
+ "description": "Time of the show",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "show_date": {
+ "description": "Date of the show",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Genre of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Address of the theatre",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "location": {
+ "description": "City where the theatre is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "movie_name": {
+ "description": "Name of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "music_1": {
+ "description": "A popular provider of a wide range of music content for searching and listening",
+ "slots": {
+ "song_name": {
+ "description": "Name of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "artist": {
+ "description": "Artist who performed the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "album": {
+ "description": "Album the song belongs to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Genre of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "year": {
+ "description": "Year in which the song was released",
+ "is_categorical": true,
+ "possible_values": [
+ "2010",
+ "2011",
+ "2012",
+ "2013",
+ "2014",
+ "2015",
+ "2016",
+ "2017",
+ "2018",
+ "2019"
+ ]
+ },
+ "playback_device": {
+ "description": "Playback device on which the song is to be played",
+ "is_categorical": true,
+ "possible_values": [
+ "tv",
+ "kitchen speaker",
+ "bedroom speaker"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "music_2": {
+ "description": "A widely used service for finding and playing music from a variety of genres and artists",
+ "slots": {
+ "song_name": {
+ "description": "Name of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "artist": {
+ "description": "Name of the artist the song is performed by",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "album": {
+ "description": "Album the song belongs to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Genre of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "playback_device": {
+ "description": "Playback device on which the song is to be played",
+ "is_categorical": true,
+ "possible_values": [
+ "tv",
+ "kitchen speaker",
+ "bedroom speaker"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "rentalcar_1": {
+ "description": "Car rental service with extensive coverage of locations and cars",
+ "slots": {
+ "type": {
+ "description": "Category to which rental car belongs",
+ "is_categorical": true,
+ "possible_values": [
+ "compact",
+ "standard",
+ "full-size"
+ ]
+ },
+ "car_name": {
+ "description": "Model name of rental car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_location": {
+ "description": "Location of rental car pickup",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_date": {
+ "description": "Date of rental car pickup",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_time": {
+ "description": "Time of rental car pickup",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_city": {
+ "description": "City to pick up the rental car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "dropoff_date": {
+ "description": "Date of rental car drop-off",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "total_price": {
+ "description": "Total price of car rental",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "rentalcar_2": {
+ "description": "Car rental service, available worldwide",
+ "slots": {
+ "car_type": {
+ "description": "Type of car for rental",
+ "is_categorical": true,
+ "possible_values": [
+ "compact",
+ "standard",
+ "full-size"
+ ]
+ },
+ "car_name": {
+ "description": "Name of car model",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_location": {
+ "description": "Pickup location for car rental",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_date": {
+ "description": "Date of pickup for car rental",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_time": {
+ "description": "Time of pickup for car rental",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_city": {
+ "description": "City in which to pick up rental car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "dropoff_date": {
+ "description": "End date of car rental reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "total_price": {
+ "description": "Total price of rental reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "restaurant_1": {
+ "description": "A leading provider for restaurant search and reservations",
+ "slots": {
+ "restaurant_name": {
+ "description": "Name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Date for the reservation or to find availability",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "Time for the reservation or to find availability",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "serves_alcohol": {
+ "description": "Boolean flag indicating if the restaurant serves alcohol",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "has_live_music": {
+ "description": "Boolean flag indicating if the restaurant has live music",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "phone_number": {
+ "description": "Phone number of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Address of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "party_size": {
+ "description": "Party size for a reservation",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6"
+ ]
+ },
+ "price_range": {
+ "description": "Price range for the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "inexpensive",
+ "moderate",
+ "expensive",
+ "very expensive"
+ ]
+ },
+ "city": {
+ "description": "City in which the restaurant is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "cuisine": {
+ "description": "Cuisine of food served in the restaurant",
+ "is_categorical": false,
+ "possible_values": [
+ "mexican",
+ "chinese",
+ "indian",
+ "american",
+ "italian"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "ridesharing_1": {
+ "description": "On-demand taxi calling service",
+ "slots": {
+ "destination": {
+ "description": "Destination for taxi ride",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "shared_ride": {
+ "description": "Boolean flag whether ride is shared with other passengers",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "ride_fare": {
+ "description": "Total fare for taxi ride",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "approximate_ride_duration": {
+ "description": "Approximate duration of ride to the destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_riders": {
+ "description": "Number of riders to call taxi for",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "ridesharing_2": {
+ "description": "App to book a cab to any destination",
+ "slots": {
+ "destination": {
+ "description": "Destination address or location for cab",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "ride_type": {
+ "description": "Type of cab ride",
+ "is_categorical": true,
+ "possible_values": [
+ "pool",
+ "regular",
+ "luxury"
+ ]
+ },
+ "ride_fare": {
+ "description": "Total fare for cab ride",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "wait_time": {
+ "description": "Expected waiting time for pick-up by cab",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_seats": {
+ "description": "Number of seats to reserve in the cab",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "services_1": {
+ "description": "A widely used service for finding and reserving the hair stylist of your choice",
+ "slots": {
+ "stylist_name": {
+ "description": "Name of the hair stylist/salon",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the stylist/salon",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "average_rating": {
+ "description": "Average review rating for the stylist/salon",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "is_unisex": {
+ "description": "Boolean flag indicating if the salon is unisex",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "street_address": {
+ "description": "Address of the stylist/salon",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "City where the salon is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_date": {
+ "description": "Date for the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_time": {
+ "description": "Time of the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "services_2": {
+ "description": "The go-to service for finding and booking appointments with top rated dentists",
+ "slots": {
+ "dentist_name": {
+ "description": "Name of the dentist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Phone number of the dentist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Address of the dentist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "City where the dentist is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_date": {
+ "description": "Date for the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_time": {
+ "description": "Time for the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "offers_cosmetic_services": {
+ "description": "Boolean flag indicating if the dentist offers cosmetic services",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "services_3": {
+ "description": "A popular provider for finding the right doctor for your needs. Also allows you to schedule your visit to the doctor",
+ "slots": {
+ "doctor_name": {
+ "description": "Name of the doctor or the medical practice",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Contact number for the doctor or the medical practice",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "average_rating": {
+ "description": "Average review rating of the doctor",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Address of the doctor",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "City where the doctor is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_date": {
+ "description": "Date for scheduling the appointment with the doctor",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_time": {
+ "description": "Time for the appointment with the doctor",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "Speciality of the doctor",
+ "is_categorical": true,
+ "possible_values": [
+ "gynecologist",
+ "ent specialist",
+ "ophthalmologist",
+ "general practitioner",
+ "dermatologist"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "travel_1": {
+ "description": "The biggest database of tourist attractions and points of interest",
+ "slots": {
+ "location": {
+ "description": "City or town where the attraction is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "attraction_name": {
+ "description": "Common name of the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "category": {
+ "description": "Category to which the attraction belongs",
+ "is_categorical": true,
+ "possible_values": [
+ "place of worship",
+ "theme park",
+ "museum",
+ "historical landmark",
+ "park",
+ "tourist attraction",
+ "sports venue",
+ "shopping area",
+ "performing arts venue",
+ "nature preserve"
+ ]
+ },
+ "phone_number": {
+ "description": "Phone number to contact the attraction",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "free_entry": {
+ "description": "Boolean flag indicating whether entrance to attraction is free",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "good_for_kids": {
+ "description": "Boolean flag indicating whether attraction is good for to take kids to",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "weather_1": {
+ "description": "Check the weather for any place and any date",
+ "slots": {
+ "precipitation": {
+ "description": "The possibility of rain or snow in percentage",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "humidity": {
+ "description": "Percentage humidity",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "wind": {
+ "description": "Wind speed in miles per hour",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "temperature": {
+ "description": "Temperature in Fahrenheit",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "Name of the city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Date for the weather",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "alarm_1": {
+ "description": "Manage alarms by getting and setting them easily",
+ "slots": {
+ "alarm_time": {
+ "description": "Time of the alarm",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "alarm_name": {
+ "description": "Name of the alarm",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "new_alarm_time": {
+ "description": "Time to set for the new alarm",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "new_alarm_name": {
+ "description": "Name to use for the new alarm",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "bank_2": {
+ "description": "Service to manage your bank accounts and finances",
+ "slots": {
+ "account_type": {
+ "description": "The user's account type",
+ "is_categorical": true,
+ "possible_values": [
+ "checking",
+ "savings"
+ ]
+ },
+ "recipient_account_type": {
+ "description": "The account type of the recipient to transfer the money to",
+ "is_categorical": true,
+ "possible_values": [
+ "checking",
+ "savings"
+ ]
+ },
+ "account_balance": {
+ "description": "The balance in the specified account",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "transfer_amount": {
+ "description": "The amount of money to transfer",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "recipient_name": {
+ "description": "The name of the recipient to transfer the money to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "transfer_time": {
+ "description": "Number of days for the transfer to go through",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "flight_3": {
+ "description": "Find one way and round trip flights to your favorite city",
+ "slots": {
+ "passengers": {
+ "description": "Number of passengers to find flight seats for",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "flight_class": {
+ "description": "Fare class of flight booking",
+ "is_categorical": true,
+ "possible_values": [
+ "economy",
+ "premium economy",
+ "business",
+ "first class"
+ ]
+ },
+ "origin_city": {
+ "description": "City in which the journey originates",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_city": {
+ "description": "City in which the journey ends",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "origin_airport_name": {
+ "description": "Number of the airport flying out from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_airport_name": {
+ "description": "Number of the airport flying to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "Date of departure flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "return_date": {
+ "description": "Date of return flight",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_stops": {
+ "description": "Number of stops in the itinerary",
+ "is_categorical": true,
+ "possible_values": [
+ "0",
+ "1"
+ ]
+ },
+ "outbound_departure_time": {
+ "description": "Local time of departure of flight from origin to destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "outbound_arrival_time": {
+ "description": "Local time of arrival of flight from origin to destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_arrival_time": {
+ "description": "Local time of arrival of flight from destination to origin",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_departure_time": {
+ "description": "Local time of departure of flight from destination to origin",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Price per passenger of the itinerary",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_checked_bags": {
+ "description": "Number of bags to check in",
+ "is_categorical": true,
+ "possible_values": [
+ "0",
+ "1",
+ "2"
+ ]
+ },
+ "airlines": {
+ "description": "Name of airline operating the flight",
+ "is_categorical": true,
+ "possible_values": [
+ "united airlines",
+ "american airlines",
+ "delta airlines",
+ "southwest airlines",
+ "alaska airlines",
+ "british airways",
+ "air canada",
+ "air france"
+ ]
+ },
+ "arrives_next_day": {
+ "description": "Whether the flight arrives the next day",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "hotel_4": {
+ "description": "Accommodation searching and booking portal",
+ "slots": {
+ "location": {
+ "description": "City or town where the accommodation is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_rooms": {
+ "description": "Number of rooms to reserve",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3"
+ ]
+ },
+ "check_in_date": {
+ "description": "Check in date for reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "stay_length": {
+ "description": "Length of stay in days",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "star_rating": {
+ "description": "Star rating of the accommodation",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "place_name": {
+ "description": "Name of the accommodation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "street_address": {
+ "description": "Street address of the accommodation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Contact phone number of the accommodation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price_per_night": {
+ "description": "Price per night for the stay",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "smoking_allowed": {
+ "description": "Whether or not smoking is allowed inside the place",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "media_2": {
+ "description": "The widest selection and lowest prices for movie rentals",
+ "slots": {
+ "movie_name": {
+ "description": "Name of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Main genre of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "subtitle_language": {
+ "description": "Language to use for subtitles (or None for no subtitles)",
+ "is_categorical": true,
+ "possible_values": [
+ "none",
+ "english",
+ "mandarin",
+ "spanish"
+ ]
+ },
+ "director": {
+ "description": "Name of the director of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "actors": {
+ "description": "Name of an actor starring in the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Cost of renting movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "movie_2": {
+ "description": "The definitive database to discover new movies to watch",
+ "slots": {
+ "title": {
+ "description": "Title for movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Genre of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "aggregate_rating": {
+ "description": "Aggregate user rating for movie, scale of 10",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "starring": {
+ "description": "Name of actor starring in movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "director": {
+ "description": "Name of director of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "restaurant_2": {
+ "description": "A popular restaurant search and reservation service",
+ "slots": {
+ "restaurant_name": {
+ "description": "Name of the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Tentative date of restaurant reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "Tentative time of restaurant reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "has_seating_outdoors": {
+ "description": "Whether the restaurant has outdoor seating available",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "has_vegetarian_options": {
+ "description": "Whether the restaurant has adequate vegetarian options",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "phone_number": {
+ "description": "Phone number to contact restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "rating": {
+ "description": "Average user rating for restaurant on a scale of 5",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Address of restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_seats": {
+ "description": "Number of seats to reserve at the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6"
+ ]
+ },
+ "price_range": {
+ "description": "Price range for the restaurant",
+ "is_categorical": true,
+ "possible_values": [
+ "cheap",
+ "moderate",
+ "pricey",
+ "ultra high-end"
+ ]
+ },
+ "location": {
+ "description": "City where the restaurant is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "category": {
+ "description": "The category of food offered by the restaurant",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "services_4": {
+ "description": "Discover the right therapist for you and make reservations easily",
+ "slots": {
+ "therapist_name": {
+ "description": "Name of the therapist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Contact number of the therapist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Address of the therapist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "Area where user wants to search for a therapist",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_date": {
+ "description": "Date of the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "appointment_time": {
+ "description": "Time of the appointment",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "type": {
+ "description": "Type of the therapist",
+ "is_categorical": true,
+ "possible_values": [
+ "psychologist",
+ "family counselor",
+ "psychiatrist"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "bus_3": {
+ "description": "Affordable and comfortable bus travel across the country",
+ "slots": {
+ "from_city": {
+ "description": "The city to depart from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to_city": {
+ "description": "The destination city of the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "from_station": {
+ "description": "Name of station of departure",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to_station": {
+ "description": "Name of station of arrival",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "The date of departure",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_time": {
+ "description": "The time of departure",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "Ticket price per passenger",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "additional_luggage": {
+ "description": "Whether to carry excess baggage in the bus",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "num_passengers": {
+ "description": "The number of tickets for the trip",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "category": {
+ "description": "How many stops the route has",
+ "is_categorical": true,
+ "possible_values": [
+ "direct",
+ "one-stop"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "event_3": {
+ "description": "Find and book tickets to any cultural events in your area",
+ "slots": {
+ "event_type": {
+ "description": "Type of cultural event",
+ "is_categorical": true,
+ "possible_values": [
+ "music",
+ "theater"
+ ]
+ },
+ "event_name": {
+ "description": "Name of artist or play",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date": {
+ "description": "Date of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "time": {
+ "description": "Start time of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_tickets": {
+ "description": "Number of tickets to reserve for the event",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5",
+ "6",
+ "7",
+ "8",
+ "9"
+ ]
+ },
+ "price_per_ticket": {
+ "description": "Price of each ticket",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "City where the event is taking place",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "venue": {
+ "description": "Exact venue of event",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "venue_address": {
+ "description": "Street address of event venue",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "flight_4": {
+ "description": "Find cheap flights in seconds and book flights",
+ "slots": {
+ "number_of_tickets": {
+ "description": "the number of flight tickets for the trip",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "seating_class": {
+ "description": "The cabin seat option",
+ "is_categorical": true,
+ "possible_values": [
+ "economy",
+ "premium economy",
+ "business"
+ ]
+ },
+ "origin_airport": {
+ "description": "The name of the airport or city to depart from",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "destination_airport": {
+ "description": "The name of the airport or city to arrive at",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "departure_date": {
+ "description": "Start date of the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "return_date": {
+ "description": "End date of the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "is_nonstop": {
+ "description": "Whether the flight is a direct one",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "outbound_departure_time": {
+ "description": "Departure time of the flight flying to the destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "outbound_arrival_time": {
+ "description": "Arrival time of the flight flying to the destination",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_arrival_time": {
+ "description": "Arrival time of the flight coming back from the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "inbound_departure_time": {
+ "description": "Departure time of the flight coming back from the trip",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price": {
+ "description": "The total cost of the flight tickets",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "airlines": {
+ "description": "The company that provides air transport services",
+ "is_categorical": true,
+ "possible_values": [
+ "united airlines",
+ "american airlines",
+ "delta airlines",
+ "southwest airlines",
+ "alaska airlines",
+ "british airways",
+ "air canada",
+ "air france",
+ "south african airways",
+ "lot polish airlines",
+ "latam brasil"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "home_2": {
+ "description": "Service for finding properties to buy and rent",
+ "slots": {
+ "intent": {
+ "description": "Whether to buy or rent a property",
+ "is_categorical": true,
+ "possible_values": [
+ "rent",
+ "buy"
+ ]
+ },
+ "area": {
+ "description": "City where the property is located",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "address": {
+ "description": "Street address of property",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "property_name": {
+ "description": "Name of property or apartment complex",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "phone_number": {
+ "description": "Contact number of property or apartment complex",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "has_garage": {
+ "description": "Whether the property has a garage",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "in_unit_laundry": {
+ "description": "Whether the property has in-unit laundry facilities",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "price": {
+ "description": "Sale price or per-month rent of property",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "visit_date": {
+ "description": "Date for visit to the property",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_beds": {
+ "description": "Number of bedrooms in the property",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "number_of_baths": {
+ "description": "Number of bathroom in the property",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "media_3": {
+ "description": "Enjoy instant and unlimited access to best shows, movies, comedy, sports, documentaries and more.",
+ "slots": {
+ "title": {
+ "description": "Title of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Category of the content",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "subtitle_language": {
+ "description": "Language of the subtitles",
+ "is_categorical": true,
+ "possible_values": [
+ "english",
+ "spanish",
+ "hindi",
+ "french"
+ ]
+ },
+ "starring": {
+ "description": "Celebs acting in the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "messaging_1": {
+ "description": "Connect and share locations with your contacts",
+ "slots": {
+ "location": {
+ "description": "Location to share with contact",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "contact_name": {
+ "description": "Name of contact to send to",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "movie_3": {
+ "description": "A review-aggregation website for movies and television",
+ "slots": {
+ "movie_title": {
+ "description": "Name of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Type of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "percent_rating": {
+ "description": "Average critic percentage rating",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "cast": {
+ "description": "Actors in the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "directed_by": {
+ "description": "Director of the movie",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "music_3": {
+ "description": "A free, personalized platform that plays music you'll love. Discover new music and enjoy old favorites.",
+ "slots": {
+ "track": {
+ "description": "Name of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "artist": {
+ "description": "Performer's name",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "album": {
+ "description": "Collection of the song",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "genre": {
+ "description": "Type of the music",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "year": {
+ "description": "Year when the song was first released",
+ "is_categorical": true,
+ "possible_values": [
+ "2010",
+ "2011",
+ "2012",
+ "2013",
+ "2014",
+ "2015",
+ "2016",
+ "2017",
+ "2018",
+ "2019"
+ ]
+ },
+ "device": {
+ "description": "Place or name of the media player to play the song selected",
+ "is_categorical": true,
+ "possible_values": [
+ "living room",
+ "kitchen",
+ "patio"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "payment_1": {
+ "description": "The fast, simple way to pay in apps, on the web, and in millions of stores",
+ "slots": {
+ "payment_method": {
+ "description": "The source of money used for making the payment",
+ "is_categorical": true,
+ "possible_values": [
+ "app balance",
+ "debit card",
+ "credit card"
+ ]
+ },
+ "amount": {
+ "description": "The amount of money to send or request",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "receiver": {
+ "description": "Name of the contact or account to make the transaction with",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "private_visibility": {
+ "description": "Whether the transaction is private or not",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "rentalcar_3": {
+ "description": "A leading global provider of car rental solutions",
+ "slots": {
+ "car_type": {
+ "description": "Type of the car",
+ "is_categorical": true,
+ "possible_values": [
+ "hatchback",
+ "sedan",
+ "suv"
+ ]
+ },
+ "car_name": {
+ "description": "Car model",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_location": {
+ "description": "Place to pick up the car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "start_date": {
+ "description": "The first date to start using the rental car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "pickup_time": {
+ "description": "Time for the pick-up",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "city": {
+ "description": "City where you want to rent the car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "end_date": {
+ "description": "The date to return the car",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "price_per_day": {
+ "description": "The cost for renting the car per day",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "add_insurance": {
+ "description": "Whether to purchase insurance",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ },
+ "train_1": {
+ "description": "Service to find and reserve train journeys between cities",
+ "slots": {
+ "from": {
+ "description": "Starting city for train journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to": {
+ "description": "Ending city for train journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "from_station": {
+ "description": "Name of station at starting city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "to_station": {
+ "description": "Name of station at ending city",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "date_of_journey": {
+ "description": "Date of train journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "journey_start_time": {
+ "description": "Time of start of train journey",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "total": {
+ "description": "Total price of train reservation",
+ "is_categorical": false,
+ "possible_values": []
+ },
+ "number_of_adults": {
+ "description": "Number of adults to reserve train tickets for",
+ "is_categorical": true,
+ "possible_values": [
+ "1",
+ "2",
+ "3",
+ "4",
+ "5"
+ ]
+ },
+ "class": {
+ "description": "Fare class for train reservation",
+ "is_categorical": true,
+ "possible_values": [
+ "value",
+ "flexible",
+ "business"
+ ]
+ },
+ "trip_protection": {
+ "description": "Whether to add trip protection to reservation, for a fee",
+ "is_categorical": true,
+ "possible_values": [
+ "true",
+ "false"
+ ]
+ },
+ "count": {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": false,
+ "possible_values": []
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "Inform the value for a slot."
+ },
+ "request": {
+ "description": "Request the value of a slot."
+ },
+ "confirm": {
+ "description": "Confirm the value of a slot before making a transactional service call."
+ },
+ "offer": {
+ "description": "Offer a certain value for a slot to the user."
+ },
+ "notify_success": {
+ "description": "Inform the user that their request was successful."
+ },
+ "notify_failure": {
+ "description": "Inform the user that their request failed."
+ },
+ "inform_count": {
+ "description": "Inform the number of items found that satisfy the user's request."
+ },
+ "offer_intent": {
+ "description": "Offer a new intent to the user."
+ },
+ "req_more": {
+ "description": "Asking the user if they need anything else."
+ },
+ "goodbye": {
+ "description": "End the dialogue."
+ },
+ "inform_intent": {
+ "description": "Express the desire to perform a certain task to the system."
+ },
+ "negate_intent": {
+ "description": "Negate the intent which has been offered by the system."
+ },
+ "affirm_intent": {
+ "description": "Agree to the intent which has been offered by the system."
+ },
+ "affirm": {
+ "description": "Agree to the system's proposition. "
+ },
+ "negate": {
+ "description": "Deny the system's proposal."
+ },
+ "select": {
+ "description": "Select a result being offered by the system."
+ },
+ "request_alts": {
+ "description": "Ask for more results besides the ones offered by the system."
+ },
+ "thank_you": {
+ "description": "Thank the system."
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "affirm",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "music_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "music_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "media_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "media_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "home_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "train_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "affirm_intent",
+ "domain": "media_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "goodbye",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "restaurant_1",
+ "slot": "intent",
+ "value": "findrestaurants"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "restaurant_1",
+ "slot": "intent",
+ "value": "reserverestaurant"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_1",
+ "slot": "intent",
+ "value": "playmovie"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "geteventdates"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_2",
+ "slot": "intent",
+ "value": "lookupmusic"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_2",
+ "slot": "intent",
+ "value": "playmedia"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_1",
+ "slot": "intent",
+ "value": "lookupsong"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_1",
+ "slot": "intent",
+ "value": "playsong"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_1",
+ "slot": "intent",
+ "value": "findevents"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_1",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "findevents"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "movie_1",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "movie_1",
+ "slot": "intent",
+ "value": "gettimesformovie"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "searchonewayflight"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_2",
+ "slot": "intent",
+ "value": "searchonewayflight"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "reserveonewayflight"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "searchroundtripflights"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "reserveroundtripflights"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_2",
+ "slot": "intent",
+ "value": "searchroundtripflights"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_1",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "ridesharing_2",
+ "slot": "intent",
+ "value": "getride"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "ridesharing_1",
+ "slot": "intent",
+ "value": "getride"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_1",
+ "slot": "intent",
+ "value": "getcarsavailable"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_1",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_2",
+ "slot": "intent",
+ "value": "getcarsavailable"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_2",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_2",
+ "slot": "intent",
+ "value": "findbus"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_1",
+ "slot": "intent",
+ "value": "findbus"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_2",
+ "slot": "intent",
+ "value": "bookhouse"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_2",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_1",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_2",
+ "slot": "intent",
+ "value": "findprovider"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_2",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_1",
+ "slot": "intent",
+ "value": "findprovider"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_1",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_3",
+ "slot": "intent",
+ "value": "findprovider"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_3",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "home_1",
+ "slot": "intent",
+ "value": "findapartment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "home_1",
+ "slot": "intent",
+ "value": "schedulevisit"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bank_1",
+ "slot": "intent",
+ "value": "checkbalance"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bank_1",
+ "slot": "intent",
+ "value": "transfermoney"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_2",
+ "slot": "intent",
+ "value": "searchhouse"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "calendar_1",
+ "slot": "intent",
+ "value": "getevents"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "calendar_1",
+ "slot": "intent",
+ "value": "getavailabletime"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "calendar_1",
+ "slot": "intent",
+ "value": "addevent"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_3",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_1",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_3",
+ "slot": "intent",
+ "value": "searchhotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_1",
+ "slot": "intent",
+ "value": "searchhotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "weather_1",
+ "slot": "intent",
+ "value": "getweather"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "travel_1",
+ "slot": "intent",
+ "value": "findattractions"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "restaurant_2",
+ "slot": "intent",
+ "value": "reserverestaurant"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_3",
+ "slot": "intent",
+ "value": "searchonewayflight"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_3",
+ "slot": "intent",
+ "value": "searchroundtripflights"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "alarm_1",
+ "slot": "intent",
+ "value": "getalarms"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "alarm_1",
+ "slot": "intent",
+ "value": "addalarm"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_4",
+ "slot": "intent",
+ "value": "findprovider"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "services_4",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "restaurant_2",
+ "slot": "intent",
+ "value": "findrestaurants"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bank_2",
+ "slot": "intent",
+ "value": "checkbalance"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bank_2",
+ "slot": "intent",
+ "value": "transfermoney"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "movie_2",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_2",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_2",
+ "slot": "intent",
+ "value": "rentmovie"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_4",
+ "slot": "intent",
+ "value": "searchhotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "hotel_4",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_3",
+ "slot": "intent",
+ "value": "lookupmusic"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "music_3",
+ "slot": "intent",
+ "value": "playmedia"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_3",
+ "slot": "intent",
+ "value": "findevents"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "event_3",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_4",
+ "slot": "intent",
+ "value": "searchonewayflight"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "flight_4",
+ "slot": "intent",
+ "value": "searchroundtripflights"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_3",
+ "slot": "intent",
+ "value": "getcarsavailable"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "rentalcar_3",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_3",
+ "slot": "intent",
+ "value": "findbus"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "bus_3",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "home_2",
+ "slot": "intent",
+ "value": "findhomebyarea"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "home_2",
+ "slot": "intent",
+ "value": "schedulevisit"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "movie_1",
+ "slot": "intent",
+ "value": "buymovietickets"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "payment_1",
+ "slot": "intent",
+ "value": "makepayment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "payment_1",
+ "slot": "intent",
+ "value": "requestpayment"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "train_1",
+ "slot": "intent",
+ "value": "findtrains"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "train_1",
+ "slot": "intent",
+ "value": "gettraintickets"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "movie_3",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_3",
+ "slot": "intent",
+ "value": "playmovie"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "media_3",
+ "slot": "intent",
+ "value": "findmovies"
+ },
+ {
+ "intent": "inform_intent",
+ "domain": "messaging_1",
+ "slot": "intent",
+ "value": "sharelocation"
+ },
+ {
+ "intent": "negate",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "media_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "media_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "train_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "media_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "negate_intent",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "media_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "music_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "flight_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "weather_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "music_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "travel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "ridesharing_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "ridesharing_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "media_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "movie_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "flight_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "movie_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "media_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "flight_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_failure",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "media_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "music_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "music_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "ridesharing_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "ridesharing_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "media_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "home_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "payment_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "train_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "media_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "notify_success",
+ "domain": "messaging_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "restaurant_1",
+ "slot": "intent",
+ "value": "reserverestaurant"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "event_2",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "music_2",
+ "slot": "intent",
+ "value": "playmedia"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "music_1",
+ "slot": "intent",
+ "value": "playsong"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "event_1",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "movie_1",
+ "slot": "intent",
+ "value": "buymovietickets"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "reserveonewayflight"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "flight_1",
+ "slot": "intent",
+ "value": "reserveroundtripflights"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "media_1",
+ "slot": "intent",
+ "value": "playmovie"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "rentalcar_1",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "rentalcar_2",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "bus_2",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "bus_1",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "services_2",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "services_1",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "services_3",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "home_1",
+ "slot": "intent",
+ "value": "schedulevisit"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "bank_1",
+ "slot": "intent",
+ "value": "transfermoney"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "hotel_2",
+ "slot": "intent",
+ "value": "bookhouse"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "calendar_1",
+ "slot": "intent",
+ "value": "addevent"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "hotel_3",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "hotel_1",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "alarm_1",
+ "slot": "intent",
+ "value": "addalarm"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "services_4",
+ "slot": "intent",
+ "value": "bookappointment"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "restaurant_2",
+ "slot": "intent",
+ "value": "reserverestaurant"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "bank_2",
+ "slot": "intent",
+ "value": "transfermoney"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "media_2",
+ "slot": "intent",
+ "value": "rentmovie"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "hotel_4",
+ "slot": "intent",
+ "value": "reservehotel"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "music_3",
+ "slot": "intent",
+ "value": "playmedia"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "event_3",
+ "slot": "intent",
+ "value": "buyeventtickets"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "rentalcar_3",
+ "slot": "intent",
+ "value": "reservecar"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "bus_3",
+ "slot": "intent",
+ "value": "buybusticket"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "home_2",
+ "slot": "intent",
+ "value": "schedulevisit"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "train_1",
+ "slot": "intent",
+ "value": "gettraintickets"
+ },
+ {
+ "intent": "offer_intent",
+ "domain": "media_3",
+ "slot": "intent",
+ "value": "playmovie"
+ },
+ {
+ "intent": "req_more",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "has_live_music",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "serves_alcohol",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "cuisine",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "price_range",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_1",
+ "slot": "title",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_1",
+ "slot": "directed_by",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_1",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_1",
+ "slot": "restaurant_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "venue_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "event_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "number_of_tickets",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_2",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_1",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_1",
+ "slot": "year",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "city_of_event",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "subcategory",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "address_of_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "number_of_seats",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "event_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "show_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "event_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "event_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_1",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_2",
+ "slot": "venue",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "origin_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "destination_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "origin_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "refundable",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "destination_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "outbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "origin",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "destination_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "outbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "seating_class",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "is_redeye",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "origin_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "passengers",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "airlines",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "outbound_departure_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "number_stops",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "return_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "inbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "return_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "number_stops",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_1",
+ "slot": "inbound_departure_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_2",
+ "slot": "inbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_2",
+ "slot": "number_of_seats",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_2",
+ "slot": "wait_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_2",
+ "slot": "ride_fare",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_2",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_2",
+ "slot": "ride_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_1",
+ "slot": "number_of_riders",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_1",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_1",
+ "slot": "ride_fare",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_1",
+ "slot": "approximate_ride_duration",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "ridesharing_1",
+ "slot": "shared_ride",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "pickup_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "dropoff_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "pickup_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "pickup_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "total_price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "dropoff_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "pickup_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "pickup_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "total_price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "pickup_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "origin",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "origin_station_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "destination_station_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "from_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "leaving_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "to_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "from_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "to_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "where_to",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "number_of_adults",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "check_in_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "check_out_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "has_laundry_service",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_2",
+ "slot": "total_price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "group_size",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "travelers",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "offers_cosmetic_services",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "appointment_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "appointment_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "average_rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "appointment_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "is_unisex",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "appointment_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "appointment_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "average_rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "appointment_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "number_of_beds",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "pets_allowed",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "furnished",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_1",
+ "slot": "visit_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_1",
+ "slot": "recipient_account_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_1",
+ "slot": "amount",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_1",
+ "slot": "account_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_2",
+ "slot": "dentist_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_1",
+ "slot": "stylist_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_3",
+ "slot": "doctor_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "calendar_1",
+ "slot": "event_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "calendar_1",
+ "slot": "event_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "calendar_1",
+ "slot": "event_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "calendar_1",
+ "slot": "event_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "check_in_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "check_out_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "average_rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "hotel_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "pets_welcome",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_3",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "destination",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "check_in_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "hotel_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "number_of_days",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "has_wifi",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "price_per_night",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "star_rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_1",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "departure_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_2",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "movie_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "weather_1",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "weather_1",
+ "slot": "humidity",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "weather_1",
+ "slot": "wind",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "weather_1",
+ "slot": "date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_2",
+ "slot": "song_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_2",
+ "slot": "artist",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_2",
+ "slot": "album",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_1",
+ "slot": "album",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_1",
+ "slot": "song_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_1",
+ "slot": "artist",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "pickup_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "car_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_1",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "pickup_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "car_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_2",
+ "slot": "car_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel_1",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel_1",
+ "slot": "good_for_kids",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel_1",
+ "slot": "free_entry",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "travel_1",
+ "slot": "location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "fare",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "leaving_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_1",
+ "slot": "transfers",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "restaurant_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "has_vegetarian_options",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "price_range",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "has_seating_outdoors",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant_2",
+ "slot": "rating",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "origin_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "destination_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "outbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "arrives_next_day",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "destination_airport_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "origin_airport_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "number_checked_bags",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "passengers",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "return_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "flight_class",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "inbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "alarm_1",
+ "slot": "new_alarm_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "appointment_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "appointment_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "services_4",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_2",
+ "slot": "recipient_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_2",
+ "slot": "transfer_amount",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_2",
+ "slot": "transfer_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bank_2",
+ "slot": "account_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_2",
+ "slot": "director",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_2",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_2",
+ "slot": "starring",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_2",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_2",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_2",
+ "slot": "actors",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_2",
+ "slot": "director",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "smoking_allowed",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "price_per_night",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "stay_length",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "check_in_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "hotel_4",
+ "slot": "street_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_3",
+ "slot": "number_stops",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_3",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "music_3",
+ "slot": "year",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "price_per_ticket",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "venue_address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "event_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "number_of_tickets",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "origin_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "destination_airport",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "outbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "number_of_tickets",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "seating_class",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "return_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "flight_4",
+ "slot": "inbound_arrival_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "pickup_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "end_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "start_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "price_per_day",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "add_insurance",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "to_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "departure_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "from_city",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "category",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "to_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "from_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "num_passengers",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "number_of_beds",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "in_unit_laundry",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "visit_date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "intent",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "has_garage",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "phone_number",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "number_of_baths",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "number_of_tickets",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "show_type",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "show_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_1",
+ "slot": "theater_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "property_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "home_2",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "payment_1",
+ "slot": "amount",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "payment_1",
+ "slot": "receiver",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "payment_1",
+ "slot": "payment_method",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "date_of_journey",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "from",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "to",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "from_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "to_station",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "number_of_adults",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "train_1",
+ "slot": "trip_protection",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_3",
+ "slot": "cast",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_3",
+ "slot": "directed_by",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "movie_3",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_3",
+ "slot": "genre",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_3",
+ "slot": "starring",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "media_3",
+ "slot": "title",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "departure_time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "bus_3",
+ "slot": "price",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "messaging_1",
+ "slot": "contact_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "date",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "event_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "venue",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "event_3",
+ "slot": "time",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "car_name",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "pickup_location",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "rentalcar_3",
+ "slot": "car_type",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "music_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "music_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "flight_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "media_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "weather_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "travel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "flight_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "movie_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "media_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "flight_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "home_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "train_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "movie_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "request_alts",
+ "domain": "media_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "restaurant_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "event_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "music_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "music_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "event_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "movie_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "flight_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "flight_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "rentalcar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "rentalcar_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "bus_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "bus_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "services_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "services_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "services_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "home_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "bank_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "calendar_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "weather_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "travel_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "flight_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "alarm_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "services_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "restaurant_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "bank_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "movie_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "hotel_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "music_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "event_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "flight_4",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "rentalcar_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "bus_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "home_2",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "train_1",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "select",
+ "domain": "movie_3",
+ "slot": "",
+ "value": ""
+ },
+ {
+ "intent": "thank_you",
+ "domain": "",
+ "slot": "",
+ "value": ""
+ }
+ ],
+ "state": {
+ "bank_1": {
+ "account_type": "",
+ "recipient_account_type": "",
+ "balance": "",
+ "amount": "",
+ "recipient_account_name": ""
+ },
+ "bus_1": {
+ "from_location": "",
+ "to_location": "",
+ "from_station": "",
+ "to_station": "",
+ "leaving_date": "",
+ "leaving_time": "",
+ "fare": "",
+ "travelers": "",
+ "transfers": ""
+ },
+ "bus_2": {
+ "origin": "",
+ "destination": "",
+ "origin_station_name": "",
+ "destination_station_name": "",
+ "departure_date": "",
+ "price": "",
+ "departure_time": "",
+ "group_size": "",
+ "fare_type": ""
+ },
+ "calendar_1": {
+ "event_date": "",
+ "event_time": "",
+ "event_location": "",
+ "event_name": "",
+ "available_start_time": "",
+ "available_end_time": ""
+ },
+ "event_1": {
+ "category": "",
+ "subcategory": "",
+ "event_name": "",
+ "date": "",
+ "time": "",
+ "number_of_seats": "",
+ "city_of_event": "",
+ "event_location": "",
+ "address_of_location": ""
+ },
+ "event_2": {
+ "event_type": "",
+ "category": "",
+ "event_name": "",
+ "date": "",
+ "time": "",
+ "number_of_tickets": "",
+ "city": "",
+ "venue": "",
+ "venue_address": ""
+ },
+ "flight_1": {
+ "passengers": "",
+ "seating_class": "",
+ "origin_city": "",
+ "destination_city": "",
+ "origin_airport": "",
+ "destination_airport": "",
+ "departure_date": "",
+ "return_date": "",
+ "number_stops": "",
+ "outbound_departure_time": "",
+ "outbound_arrival_time": "",
+ "inbound_arrival_time": "",
+ "inbound_departure_time": "",
+ "price": "",
+ "refundable": "",
+ "airlines": ""
+ },
+ "flight_2": {
+ "passengers": "",
+ "seating_class": "",
+ "origin": "",
+ "destination": "",
+ "origin_airport": "",
+ "destination_airport": "",
+ "departure_date": "",
+ "return_date": "",
+ "number_stops": "",
+ "outbound_departure_time": "",
+ "outbound_arrival_time": "",
+ "inbound_arrival_time": "",
+ "inbound_departure_time": "",
+ "fare": "",
+ "is_redeye": "",
+ "airlines": ""
+ },
+ "home_1": {
+ "area": "",
+ "address": "",
+ "property_name": "",
+ "phone_number": "",
+ "furnished": "",
+ "pets_allowed": "",
+ "rent": "",
+ "visit_date": "",
+ "number_of_beds": "",
+ "number_of_baths": ""
+ },
+ "hotel_1": {
+ "destination": "",
+ "number_of_rooms": "",
+ "check_in_date": "",
+ "number_of_days": "",
+ "star_rating": "",
+ "hotel_name": "",
+ "street_address": "",
+ "phone_number": "",
+ "price_per_night": "",
+ "has_wifi": ""
+ },
+ "hotel_2": {
+ "where_to": "",
+ "number_of_adults": "",
+ "check_in_date": "",
+ "check_out_date": "",
+ "rating": "",
+ "address": "",
+ "phone_number": "",
+ "total_price": "",
+ "has_laundry_service": ""
+ },
+ "hotel_3": {
+ "location": "",
+ "number_of_rooms": "",
+ "check_in_date": "",
+ "check_out_date": "",
+ "average_rating": "",
+ "hotel_name": "",
+ "street_address": "",
+ "phone_number": "",
+ "price": "",
+ "pets_welcome": ""
+ },
+ "media_1": {
+ "title": "",
+ "genre": "",
+ "subtitles": "",
+ "directed_by": ""
+ },
+ "movie_1": {
+ "price": "",
+ "number_of_tickets": "",
+ "show_type": "",
+ "theater_name": "",
+ "show_time": "",
+ "show_date": "",
+ "genre": "",
+ "street_address": "",
+ "location": "",
+ "movie_name": ""
+ },
+ "music_1": {
+ "song_name": "",
+ "artist": "",
+ "album": "",
+ "genre": "",
+ "year": "",
+ "playback_device": ""
+ },
+ "music_2": {
+ "song_name": "",
+ "artist": "",
+ "album": "",
+ "genre": "",
+ "playback_device": ""
+ },
+ "rentalcar_1": {
+ "type": "",
+ "car_name": "",
+ "pickup_location": "",
+ "pickup_date": "",
+ "pickup_time": "",
+ "pickup_city": "",
+ "dropoff_date": "",
+ "total_price": ""
+ },
+ "rentalcar_2": {
+ "car_type": "",
+ "car_name": "",
+ "pickup_location": "",
+ "pickup_date": "",
+ "pickup_time": "",
+ "pickup_city": "",
+ "dropoff_date": "",
+ "total_price": ""
+ },
+ "restaurant_1": {
+ "restaurant_name": "",
+ "date": "",
+ "time": "",
+ "serves_alcohol": "",
+ "has_live_music": "",
+ "phone_number": "",
+ "street_address": "",
+ "party_size": "",
+ "price_range": "",
+ "city": "",
+ "cuisine": ""
+ },
+ "ridesharing_1": {
+ "destination": "",
+ "shared_ride": "",
+ "ride_fare": "",
+ "approximate_ride_duration": "",
+ "number_of_riders": ""
+ },
+ "ridesharing_2": {
+ "destination": "",
+ "ride_type": "",
+ "ride_fare": "",
+ "wait_time": "",
+ "number_of_seats": ""
+ },
+ "services_1": {
+ "stylist_name": "",
+ "phone_number": "",
+ "average_rating": "",
+ "is_unisex": "",
+ "street_address": "",
+ "city": "",
+ "appointment_date": "",
+ "appointment_time": ""
+ },
+ "services_2": {
+ "dentist_name": "",
+ "phone_number": "",
+ "address": "",
+ "city": "",
+ "appointment_date": "",
+ "appointment_time": "",
+ "offers_cosmetic_services": ""
+ },
+ "services_3": {
+ "doctor_name": "",
+ "phone_number": "",
+ "average_rating": "",
+ "street_address": "",
+ "city": "",
+ "appointment_date": "",
+ "appointment_time": "",
+ "type": ""
+ },
+ "travel_1": {
+ "location": "",
+ "attraction_name": "",
+ "category": "",
+ "phone_number": "",
+ "free_entry": "",
+ "good_for_kids": ""
+ },
+ "weather_1": {
+ "precipitation": "",
+ "humidity": "",
+ "wind": "",
+ "temperature": "",
+ "city": "",
+ "date": ""
+ },
+ "alarm_1": {
+ "alarm_time": "",
+ "alarm_name": "",
+ "new_alarm_time": "",
+ "new_alarm_name": ""
+ },
+ "bank_2": {
+ "account_type": "",
+ "recipient_account_type": "",
+ "account_balance": "",
+ "transfer_amount": "",
+ "recipient_name": "",
+ "transfer_time": ""
+ },
+ "flight_3": {
+ "passengers": "",
+ "flight_class": "",
+ "origin_city": "",
+ "destination_city": "",
+ "origin_airport_name": "",
+ "destination_airport_name": "",
+ "departure_date": "",
+ "return_date": "",
+ "number_stops": "",
+ "outbound_departure_time": "",
+ "outbound_arrival_time": "",
+ "inbound_arrival_time": "",
+ "inbound_departure_time": "",
+ "price": "",
+ "number_checked_bags": "",
+ "airlines": "",
+ "arrives_next_day": ""
+ },
+ "hotel_4": {
+ "location": "",
+ "number_of_rooms": "",
+ "check_in_date": "",
+ "stay_length": "",
+ "star_rating": "",
+ "place_name": "",
+ "street_address": "",
+ "phone_number": "",
+ "price_per_night": "",
+ "smoking_allowed": ""
+ },
+ "media_2": {
+ "movie_name": "",
+ "genre": "",
+ "subtitle_language": "",
+ "director": "",
+ "actors": "",
+ "price": ""
+ },
+ "movie_2": {
+ "title": "",
+ "genre": "",
+ "aggregate_rating": "",
+ "starring": "",
+ "director": ""
+ },
+ "restaurant_2": {
+ "restaurant_name": "",
+ "date": "",
+ "time": "",
+ "has_seating_outdoors": "",
+ "has_vegetarian_options": "",
+ "phone_number": "",
+ "rating": "",
+ "address": "",
+ "number_of_seats": "",
+ "price_range": "",
+ "location": "",
+ "category": ""
+ },
+ "services_4": {
+ "therapist_name": "",
+ "phone_number": "",
+ "address": "",
+ "city": "",
+ "appointment_date": "",
+ "appointment_time": "",
+ "type": ""
+ },
+ "bus_3": {
+ "from_city": "",
+ "to_city": "",
+ "from_station": "",
+ "to_station": "",
+ "departure_date": "",
+ "departure_time": "",
+ "price": "",
+ "additional_luggage": "",
+ "num_passengers": "",
+ "category": ""
+ },
+ "event_3": {
+ "event_type": "",
+ "event_name": "",
+ "date": "",
+ "time": "",
+ "number_of_tickets": "",
+ "price_per_ticket": "",
+ "city": "",
+ "venue": "",
+ "venue_address": ""
+ },
+ "flight_4": {
+ "number_of_tickets": "",
+ "seating_class": "",
+ "origin_airport": "",
+ "destination_airport": "",
+ "departure_date": "",
+ "return_date": "",
+ "is_nonstop": "",
+ "outbound_departure_time": "",
+ "outbound_arrival_time": "",
+ "inbound_arrival_time": "",
+ "inbound_departure_time": "",
+ "price": "",
+ "airlines": ""
+ },
+ "home_2": {
+ "intent": "",
+ "area": "",
+ "address": "",
+ "property_name": "",
+ "phone_number": "",
+ "has_garage": "",
+ "in_unit_laundry": "",
+ "price": "",
+ "visit_date": "",
+ "number_of_beds": "",
+ "number_of_baths": ""
+ },
+ "media_3": {
+ "title": "",
+ "genre": "",
+ "subtitle_language": "",
+ "starring": ""
+ },
+ "messaging_1": {
+ "location": "",
+ "contact_name": ""
+ },
+ "movie_3": {
+ "movie_title": "",
+ "genre": "",
+ "percent_rating": "",
+ "cast": "",
+ "directed_by": ""
+ },
+ "music_3": {
+ "track": "",
+ "artist": "",
+ "album": "",
+ "genre": "",
+ "year": "",
+ "device": ""
+ },
+ "payment_1": {
+ "payment_method": "",
+ "amount": "",
+ "receiver": "",
+ "private_visibility": ""
+ },
+ "rentalcar_3": {
+ "car_type": "",
+ "car_name": "",
+ "pickup_location": "",
+ "start_date": "",
+ "pickup_time": "",
+ "city": "",
+ "end_date": "",
+ "price_per_day": "",
+ "add_insurance": ""
+ },
+ "train_1": {
+ "from": "",
+ "to": "",
+ "from_station": "",
+ "to_station": "",
+ "date_of_journey": "",
+ "journey_start_time": "",
+ "total": "",
+ "number_of_adults": "",
+ "class": "",
+ "trip_protection": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/schema/original_data.zip b/data/unified_datasets/schema/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..110e958b8b1f73f0c57f8660f9201a2dfb340f71
Binary files /dev/null and b/data/unified_datasets/schema/original_data.zip differ
diff --git a/data/unified_datasets/schema/preprocess.py b/data/unified_datasets/schema/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..46913390c7c89978965b0c1ae65067b4c14fe9a9
--- /dev/null
+++ b/data/unified_datasets/schema/preprocess.py
@@ -0,0 +1,549 @@
+import zipfile
+import json
+import os
+from pprint import pprint
+from copy import deepcopy
+from collections import Counter
+from tqdm import tqdm
+import numpy as np
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+import re
+self_dir = os.path.dirname(os.path.abspath(__file__))
+
+
+norm_service2domain = {
+ 'alarm': 'alarm',
+ 'banks': 'bank',
+ 'buses': 'bus',
+ 'calendar': 'calendar',
+ 'events': 'event',
+ 'flights': 'flight',
+ 'homes': 'home',
+ 'hotels': 'hotel',
+ 'media': 'media',
+ 'messaging': 'messaging',
+ 'movies': 'movie',
+ 'music': 'music',
+ 'payment': 'payment',
+ 'rentalcars': 'rentalcar',
+ 'restaurants': 'restaurant',
+ 'ridesharing': 'ridesharing',
+ 'services': 'services',
+ 'trains': 'train',
+ 'travel': 'travel',
+ 'weather': 'weather'
+}
+
+digit2word = {
+ '0': 'zero', '1': 'one', '2': 'two', '3': 'three', '4': 'four', '5': 'five',
+ '6': 'six', '7': 'seven', '8': 'eight', '9': 'nine', '10': 'ten'
+}
+
+match = {
+ '0': 0,
+ '1': 0,
+ '>1': 0,
+}
+
+
+def service2domain(service):
+ s, i = service.split('_')
+ return norm_service2domain[s.lower()]+'_'+i
+
+
+def slot_normalize(service, slot):
+ pass
+
+
+def pharse_in_sen(phrase, sen):
+ '''
+ match value in the sentence
+ :param phrase: str
+ :param sen: str
+ :return: start, end if matched, else None, None
+ '''
+ assert isinstance(phrase, str)
+ pw = '(^|[\s,\.:\?!-])(?P<v>{})([\s,\.:\?!-]|$)'
+ pn = '(^|[\s\?!-]|\D[,\.:])(?P<v>{})($|[\s\?!-]|[,\.:]\D|[,\.:]$)'
+ if phrase.isdigit():
+ pattern = pn
+ else:
+ pattern = pw
+ p = re.compile(pattern.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ if num > 1:
+ match['>1'] += 1
+ else:
+ match['1'] += 1
+ return m.span('v'), num
+ if phrase.isdigit() and phrase in digit2word:
+ phrase = digit2word[phrase]
+ p = re.compile(pw.format(re.escape(phrase)), re.I)
+ m = re.search(p, sen)
+ if m:
+ num = len(re.findall(p, sen))
+ if num > 1:
+ match['>1'] += 1
+ else:
+ match['1'] += 1
+ return m.span('v'), num
+ match['0'] += 1
+ return (None, None), 0
+
+
+def number_in_sen(word, sen):
+ if ' '+word+' ' in sen:
+ return sen.index(' ' + word + ' ') + 1, sen.index(' ' + word + ' ') + 1 + len(word)
+ elif ' '+word+'.' in sen:
+ return sen.index(' ' + word + '.') + 1, sen.index(' ' + word + '.') + 1 + len(word)
+ elif ' '+word+',' in sen:
+ return sen.index(' ' + word + ',') + 1, sen.index(' ' + word + ',') + 1 + len(word)
+ elif sen.startswith(word+ ' ') or sen.startswith(word+'.') or sen.startswith(word+','):
+ return 0, len(word)
+ elif word.isdigit() and word in digit2word:
+ ori_word = word
+ ori_sen = sen
+ word = digit2word[word]
+ sen = sen.lower()
+ if ' ' + word + ' ' in sen:
+ return sen.index(' ' + word + ' ') + 1, sen.index(' ' + word + ' ') + 1 + len(word)
+ elif ' ' + word + '.' in sen:
+ return sen.index(' ' + word + '.') + 1, sen.index(' ' + word + '.') + 1 + len(word)
+ elif ' ' + word + ',' in sen:
+ return sen.index(' ' + word + ',') + 1, sen.index(' ' + word + ',') + 1 + len(word)
+ elif sen.startswith(word + ' ') or sen.startswith(word + '.') or sen.startswith(word + ','):
+ return 0, len(word)
+ word = ori_word
+ sen = ori_sen
+ return sen.index(word)
+
+
+def sys_intent():
+ return {
+ "inform": {"description": "Inform the value for a slot to the user."},
+ "request": {"description": "Request the value of a slot from the user."},
+ "confirm": {"description": "Confirm the value of a slot before making a transactional service call."},
+ "offer": {"description": "Offer a certain value for a slot to the user."},
+ "notify_success": {"description": "Inform the user that their request was successful."},
+ "notify_failure": {"description": "Inform the user that their request failed."},
+ "inform_count": {"description": "Inform the number of items found that satisfy the user's request."},
+ "offer_intent": {"description": "Offer a new intent to the user."},
+ "req_more": {"description": "Asking the user if they need anything else."},
+ "goodbye": {"description": "End the dialogue."},
+ }
+
+
+def usr_intent():
+ return {
+ "inform_intent": {"description": "Express the desire to perform a certain task to the system."},
+ "negate_intent": {"description": "Negate the intent which has been offered by the system."},
+ "affirm_intent": {"description": "Agree to the intent which has been offered by the system."},
+ "inform": {"description": "Inform the value of a slot to the system."},
+ "request": {"description": "Request the value of a slot from the system."},
+ "affirm": {"description": "Agree to the system's proposition. "},
+ "negate": {"description": "Deny the system's proposal."},
+ "select": {"description": "Select a result being offered by the system."},
+ "request_alts": {"description": "Ask for more results besides the ones offered by the system."},
+ "thank_you": {"description": "Thank the system."},
+ "goodbye": {"description": "End the dialogue."},
+ }
+
+
+def get_intent():
+ """merge sys & usr intent"""
+ return {
+ "inform": {"description": "Inform the value for a slot."},
+ "request": {"description": "Request the value of a slot."},
+ "confirm": {"description": "Confirm the value of a slot before making a transactional service call."},
+ "offer": {"description": "Offer a certain value for a slot to the user."},
+ "notify_success": {"description": "Inform the user that their request was successful."},
+ "notify_failure": {"description": "Inform the user that their request failed."},
+ "inform_count": {"description": "Inform the number of items found that satisfy the user's request."},
+ "offer_intent": {"description": "Offer a new intent to the user."},
+ "req_more": {"description": "Asking the user if they need anything else."},
+ "goodbye": {"description": "End the dialogue."},
+ "inform_intent": {"description": "Express the desire to perform a certain task to the system."},
+ "negate_intent": {"description": "Negate the intent which has been offered by the system."},
+ "affirm_intent": {"description": "Agree to the intent which has been offered by the system."},
+ "affirm": {"description": "Agree to the system's proposition. "},
+ "negate": {"description": "Deny the system's proposal."},
+ "select": {"description": "Select a result being offered by the system."},
+ "request_alts": {"description": "Ask for more results besides the ones offered by the system."},
+ "thank_you": {"description": "Thank the system."},
+ }
+
+
+def preprocess():
+ processed_dialogue = []
+ ontology = {'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': [],
+ 'state': {}}
+ ontology['intents'].update(get_intent())
+ numerical_slots = {}
+ original_zipped_path = os.path.join(self_dir, 'original_data.zip')
+ new_dir = os.path.join(self_dir, 'original_data')
+ if not os.path.exists(original_zipped_path):
+ raise FileNotFoundError(original_zipped_path)
+ if not os.path.exists(os.path.join(self_dir, 'data.zip')) or not os.path.exists(os.path.join(self_dir, 'ontology.json')):
+ print('unzip to', new_dir)
+ print('This may take several minutes')
+ archive = zipfile.ZipFile(original_zipped_path, 'r')
+ archive.extractall(self_dir)
+ cnt = 1
+ non_cate_slot_update_cnt = 0
+ non_cate_slot_update_fail_cnt = 0
+ state_cnt = {}
+ num_train_dialog = 0
+ num_train_utt = 0
+ for data_split in ['train', 'dev', 'test']:
+ dataset_name = 'schema'
+ data_dir = os.path.join(new_dir, data_split)
+ # schema => ontology
+ f = open(os.path.join(data_dir, 'schema.json'))
+ data = json.load(f)
+ for schema in data:
+ domain = service2domain(schema['service_name'])
+ ontology['domains'].setdefault(domain, {})
+ ontology['domains'][domain]['description'] = schema['description']
+ ontology['domains'][domain].setdefault('slots', {})
+ ontology['state'].setdefault(domain, {})
+ for slot in schema['slots']:
+ # numerical => non-categorical: not use
+ # is_numerical = slot['is_categorical']
+ # for value in slot['possible_values']:
+ # if not value.isdigit():
+ # is_numerical = False
+ # break
+ # if is_numerical:
+ # numerical_slots.setdefault(slot['name'].lower(), 1)
+ lower_values = [x.lower() for x in slot['possible_values']]
+ ontology['domains'][domain]['slots'][slot['name'].lower()] = {
+ "description": slot['description'],
+ "is_categorical": slot['is_categorical'],
+ "possible_values": lower_values
+ }
+ ontology['state'][domain][slot['name'].lower()] = ''
+ # add 'count' slot
+ ontology['domains'][domain]['slots']['count'] = {
+ "description": "the number of items found that satisfy the user's request.",
+ "is_categorical": False,
+ "possible_values": []
+ }
+ # ontology['state'][domain]['count'] = ''
+ # pprint(numerical_slots)
+ # dialog
+ for root, dirs, files in os.walk(data_dir):
+ fs = sorted([x for x in files if 'dialogues' in x])
+ for f in tqdm(fs, desc='processing schema-guided-{}'.format(data_split)):
+ data = json.load(open(os.path.join(data_dir, f)))
+ if data_split == 'train':
+ num_train_dialog += len(data)
+ for d in data:
+ dialogue = {
+ "dataset": dataset_name,
+ "data_split": data_split if data_split!='dev' else 'val',
+ "dialogue_id": dataset_name+'_'+str(cnt),
+ "original_id": d['dialogue_id'],
+ "domains": [service2domain(s) for s in d['services']],
+ "turns": []
+ }
+ # if d['dialogue_id'] != '84_00008':
+ # continue
+ cnt += 1
+ prev_sys_frames = []
+ prev_user_frames = []
+ all_slot_spans_from_da = []
+ state = {}
+ for domain in dialogue['domains']:
+ state.setdefault(domain, deepcopy(ontology['state'][domain]))
+ if data_split == 'train':
+ num_train_utt += len(d['turns'])
+ for utt_idx, t in enumerate(d['turns']):
+ speaker = t['speaker'].lower()
+ turn = {
+ 'speaker': speaker,
+ 'utterance': t['utterance'],
+ 'utt_idx': utt_idx,
+ 'dialogue_act': {
+ 'binary': [],
+ 'categorical': [],
+ 'non-categorical': [],
+ },
+ }
+ for i, frame in enumerate(t['frames']):
+ domain = service2domain(frame['service'])
+ for action in frame['actions']:
+ intent = action['act'].lower()
+ assert intent in ontology['intents'], [intent]
+ slot = action['slot'].lower()
+ value_list = action['values']
+ if action['act'] in ['REQ_MORE', 'AFFIRM', 'NEGATE', 'THANK_YOU', 'GOODBYE']:
+ turn['dialogue_act']['binary'].append({
+ "intent": intent,
+ "domain": '',
+ "slot": '',
+ "value": '',
+ })
+ elif action['act'] in ['NOTIFY_SUCCESS', 'NOTIFY_FAILURE', 'REQUEST_ALTS', 'AFFIRM_INTENT', 'NEGATE_INTENT']:
+ # Slot and values are always empty
+ turn['dialogue_act']['binary'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": '',
+ "value": '',
+ })
+ elif action['act'] in ['OFFER_INTENT', 'INFORM_INTENT']:
+ # always has "intent" as the slot, and a single value containing the intent being offered.
+ assert slot == 'intent'
+ turn['dialogue_act']['binary'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot,
+ "value": value_list[0].lower(),
+ })
+ elif action['act'] in ['REQUEST', 'SELECT'] and not value_list:
+ # always contains a slot, but values are optional.
+ # assert slot in ontology['domains'][domain]['slots']
+ turn['dialogue_act']['binary'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot,
+ "value": '',
+ })
+ elif action['act'] in ['INFORM_COUNT']:
+ # always has "count" as the slot, and a single element in values for the number of results obtained by the system.
+ value = value_list[0]
+ assert slot in ontology['domains'][domain]['slots']
+ (start, end), num = pharse_in_sen(value, t['utterance'])
+ if num:
+ assert value.lower() == t['utterance'][start:end].lower() \
+ or digit2word[value].lower() == t['utterance'][start:end].lower()
+ turn['dialogue_act']['non-categorical'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot.lower(),
+ "value": t['utterance'][start:end].lower(),
+ "start": start,
+ "end": end
+ })
+ else:
+ # have slot & value
+ if ontology['domains'][domain]['slots'][slot]['is_categorical']:
+ for value in value_list:
+ value = value.lower()
+ if value not in ontology['domains'][domain]['slots'][slot]['possible_values'] and value != 'dontcare':
+ ontology['domains'][domain]['slots'][slot]['possible_values'].append(value)
+ print('add value to ontology', domain, slot, value)
+ assert value in ontology['domains'][domain]['slots'][slot][
+ 'possible_values'] or value == 'dontcare'
+ turn['dialogue_act']['categorical'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot,
+ "value": value,
+ })
+ elif slot in numerical_slots:
+ value = value_list[-1]
+ (start, end), num = pharse_in_sen(value, t['utterance'])
+ if num:
+ assert value.lower() == t['utterance'][start:end].lower() \
+ or digit2word[value].lower() == t['utterance'][start:end].lower()
+ turn['dialogue_act']['non-categorical'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot.lower(),
+ "value": t['utterance'][start:end].lower(),
+ "start": start,
+ "end": end
+ })
+ else:
+ # span info in frame['slots']
+ for value in value_list:
+ for slot_info in frame['slots']:
+ start = slot_info['start']
+ end = slot_info['exclusive_end']
+ if slot_info['slot'] == slot and t['utterance'][start:end] == value:
+ turn['dialogue_act']['non-categorical'].append({
+ "intent": intent,
+ "domain": domain,
+ "slot": slot,
+ "value": value.lower(),
+ "start": start,
+ "end": end
+ })
+ break
+ # add span da to all_slot_spans_from_da
+ for ele in turn['dialogue_act']['non-categorical']:
+ all_slot_spans_from_da.insert(0, {
+ "domain": ele["domain"],
+ "slot": ele["slot"],
+ "value": ele["value"].lower(),
+ "utt_idx": utt_idx,
+ "start": ele["start"],
+ "end": ele["end"]
+ })
+ if speaker == 'user':
+ # DONE: record state update, may come from sys acts
+ # prev_state: state. update the state using current frames.
+ # candidate span info from prev frames and current frames
+ slot_spans = []
+ for frame in t['frames']:
+ for ele in frame['slots']:
+ slot, start, end = ele['slot'].lower(), ele['start'], ele['exclusive_end']
+ slot_spans.append({
+ "domain": service2domain(frame['service']),
+ "slot": slot,
+ "value": t['utterance'][start:end].lower(),
+ "utt_idx": utt_idx,
+ "start": start,
+ "end": end
+ })
+ for frame in prev_sys_frames:
+ for ele in frame['slots']:
+ slot, start, end = ele['slot'].lower(), ele['start'], ele['exclusive_end']
+ slot_spans.append({
+ "domain": service2domain(frame['service']),
+ "slot": slot,
+ "value": d['turns'][utt_idx-1]['utterance'][start:end].lower(),
+ "utt_idx": utt_idx-1,
+ "start": start,
+ "end": end
+ })
+ # turn['slot_spans'] = slot_spans
+ # turn['all_slot_span'] = deepcopy(all_slot_spans_from_da)
+ state_update = {"categorical": [], "non-categorical": []}
+ # print(utt_idx)
+ for frame in t['frames']:
+ domain = service2domain(frame['service'])
+ # print(domain)
+ for slot, value_list in frame['state']['slot_values'].items():
+ # For categorical slots, this list contains a single value assigned to the slot.
+ # For non-categorical slots, all the values in this list are spoken variations
+ # of each other and are equivalent (e.g, "6 pm", "six in the evening",
+ # "evening at 6" etc.).
+ numerical_equal_values = []
+ if slot in numerical_slots:
+ for value in value_list:
+ if value in digit2word:
+ numerical_equal_values.append(digit2word[value])
+ value_list += numerical_equal_values
+ assert len(value_list) > 0, print(slot, value_list)
+ assert slot in state[domain]
+ value_list = list(set([x.lower() for x in value_list]))
+ if state[domain][slot] in value_list:
+ continue
+ # new value
+ candidate_values = value_list
+ for prev_user_frame in prev_user_frames:
+ prev_domain = service2domain(prev_user_frame['service'])
+ if prev_domain == domain and slot in prev_user_frame['state']['slot_values']:
+ prev_value_list = [x.lower() for x in prev_user_frame['state']['slot_values'][slot]]
+ candidate_values = list(set(value_list) - set(prev_value_list))
+ assert state[domain][slot] not in candidate_values
+ assert candidate_values
+
+ if ontology['domains'][domain]['slots'][slot]['is_categorical']:
+ state_cnt.setdefault('cate_slot_update', 0)
+ state_cnt['cate_slot_update'] += 1
+ value = candidate_values[0]
+ state_update['categorical'].append(
+ {"domain": domain, "slot": slot, "value": value}
+ )
+ state[domain][slot] = value
+ else:
+ state_cnt.setdefault('non_cate_slot_update', 0)
+ state_cnt['non_cate_slot_update'] += 1
+ span_priority = []
+ slot_spans_len = len(slot_spans)
+ all_slot_spans = slot_spans+all_slot_spans_from_da
+ for span_idx, slot_span in enumerate(all_slot_spans):
+ priority = 0
+ span_domain = slot_span['domain']
+ span_slot = slot_span['slot']
+ span_value = slot_span['value']
+ if domain == span_domain:
+ priority += 1
+ if slot == span_slot:
+ priority += 10
+ if span_value in candidate_values:
+ priority += 100
+ if span_idx + 1 <= slot_spans_len:
+ priority += 0.5
+ span_priority.append(priority)
+ if span_idx + 1 <= slot_spans_len:
+ # slot_spans not run out
+ if max(span_priority) >= 111.5:
+ break
+ else:
+ # search in previous da
+ if max(span_priority) >= 111:
+ break
+ if span_priority and max(span_priority) >= 100:
+ # {111.5: 114255,
+ # 111: 29591,
+ # 100: 15208,
+ # 110: 2159,
+ # 100.5: 642,
+ # 110.5: 125,
+ # 101: 24}
+ max_priority = max(span_priority)
+ state_cnt.setdefault('max_priority', Counter())
+ state_cnt['max_priority'][max_priority] += 1
+ span_idx = np.argmax(span_priority)
+ ele = all_slot_spans[span_idx]
+ state_update['non-categorical'].append({
+ "domain": domain,
+ "slot": slot,
+ "value": ele['value'],
+ "utt_idx": ele["utt_idx"],
+ "start": ele["start"],
+ "end": ele["end"]
+ })
+ state[domain][slot] = ele['value']
+ else:
+ # not found
+ value = candidate_values[0]
+ state_update['non-categorical'].append(
+ {"domain": domain, "slot": slot, "value": value}
+ )
+ state[domain][slot] = value
+ # print(t['utterance'])
+ non_cate_slot_update_fail_cnt += 1
+ non_cate_slot_update_cnt += 1
+ turn['state'] = deepcopy(state)
+ turn['state_update'] = state_update
+ prev_user_frames = deepcopy(t['frames'])
+ else:
+ prev_sys_frames = deepcopy(t['frames'])
+
+ for da in turn['dialogue_act']['binary']:
+ if da not in ontology['binary_dialogue_act']:
+ ontology['binary_dialogue_act'].append(deepcopy(da))
+ dialogue['turns'].append(deepcopy(turn))
+ assert len(dialogue['turns']) % 2 == 0
+ dialogue['turns'].pop()
+ processed_dialogue.append(dialogue)
+ # break
+ # sort ontology binary
+ pprint(state_cnt)
+ ontology['binary_dialogue_act'] = sorted(ontology['binary_dialogue_act'], key=lambda x:x['intent'])
+ json.dump(ontology, open(os.path.join(self_dir, 'ontology.json'), 'w'), indent=2)
+ json.dump(processed_dialogue, open('data.json', 'w'), indent=2)
+ write_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ os.remove('data.json')
+ print('# train dialog: {}, # train utterance: {}'.format(num_train_dialog, num_train_utt))
+ print(non_cate_slot_update_fail_cnt, non_cate_slot_update_cnt) # 395 162399
+
+ else:
+ # read from file
+ processed_dialogue = read_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ ontology = json.load(open(os.path.join(self_dir, 'ontology.json')))
+ return processed_dialogue, ontology
+
+
+if __name__ == '__main__':
+ preprocess()
+ print(match) # {'0': 4146, '1': 53626, '>1': 2904} =>(after user act released) {'0': 487, '1': 63886, '>1': 3097}
diff --git a/data/unified_datasets/taskmaster/README.md b/data/unified_datasets/taskmaster/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..303a82317a50d1aa4f40bd625ad17ccf97da392a
--- /dev/null
+++ b/data/unified_datasets/taskmaster/README.md
@@ -0,0 +1,26 @@
+# README
+
+## Features
+
+- Annotations: character-level span for non-categorical slots. No slot descriptions.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 30483 | 540311 | 17.72 | 9.18 | 13 |
+
+## Main changes
+
+- each speaker for one turn
+- intent is set to **inform**
+- not annotate state and state upadte
+- span info is provided by original data
+
+## Original data
+
+https://github.com/google-research-datasets/Taskmaster
+
+TM-1: https://github.com/google-research-datasets/Taskmaster/tree/master/TM-1-2019
+
+TM-2: https://github.com/google-research-datasets/Taskmaster/tree/master/TM-2-2020
\ No newline at end of file
diff --git a/data/unified_datasets/taskmaster/data.zip b/data/unified_datasets/taskmaster/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..f52a3808df8a413962c71f305c64d437eb196e00
Binary files /dev/null and b/data/unified_datasets/taskmaster/data.zip differ
diff --git a/data/unified_datasets/taskmaster/ontology.json b/data/unified_datasets/taskmaster/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..9b5532e2bbcc36039486954e5b7bee1205a33cb5
--- /dev/null
+++ b/data/unified_datasets/taskmaster/ontology.json
@@ -0,0 +1,1168 @@
+{
+ "domains": {
+ "uber_lyft": {
+ "description": "order a car for a ride inside a city",
+ "slots": {
+ "location.from": {
+ "description": "pickup location",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5764,
+ "in original ontology": true
+ },
+ "location.to": {
+ "description": "destination of the ride",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6026,
+ "in original ontology": true
+ },
+ "type.ride": {
+ "description": "type of ride",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 13317,
+ "in original ontology": true
+ },
+ "num.people": {
+ "description": "number of people",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2636,
+ "in original ontology": true
+ },
+ "price.estimate": {
+ "description": "estimated cost of the ride",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4996,
+ "in original ontology": true
+ },
+ "duration.estimate": {
+ "description": "estimated duration of the ride",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1152,
+ "in original ontology": true
+ },
+ "time.pickup": {
+ "description": "time of pickup",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4303,
+ "in original ontology": true
+ },
+ "time.dropoff": {
+ "description": "time of dropoff",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 111,
+ "in original ontology": true
+ }
+ }
+ },
+ "movie_ticket": {
+ "description": "book movie tickets for a film",
+ "slots": {
+ "name.movie": {
+ "description": "name of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 8959,
+ "in original ontology": true
+ },
+ "name.theater": {
+ "description": "name of the theater",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6842,
+ "in original ontology": true
+ },
+ "num.tickets": {
+ "description": "number of tickets",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 7368,
+ "in original ontology": true
+ },
+ "time.start": {
+ "description": "start time of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 14820,
+ "in original ontology": true
+ },
+ "location.theater": {
+ "description": "location of the theater",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5295,
+ "in original ontology": true
+ },
+ "price.ticket": {
+ "description": "price of the ticket",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2418,
+ "in original ontology": true
+ },
+ "type.screening": {
+ "description": "type of the screening",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4579,
+ "in original ontology": true
+ },
+ "time.end": {
+ "description": "end time of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 10,
+ "in original ontology": true
+ },
+ "time.duration": {
+ "description": "duration of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 250,
+ "in original ontology": true
+ }
+ }
+ },
+ "restaurant_reservation": {
+ "description": "searching for a restaurant and make reservation",
+ "slots": {
+ "name.restaurant": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 14544,
+ "in original ontology": true
+ },
+ "name.reservation": {
+ "description": "name of the person who make the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 609,
+ "in original ontology": true
+ },
+ "num.guests": {
+ "description": "number of guests",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6962,
+ "in original ontology": true
+ },
+ "time.reservation": {
+ "description": "time of the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 12073,
+ "in original ontology": true
+ },
+ "type.seating": {
+ "description": "type of the seating",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 8983,
+ "in original ontology": true
+ },
+ "location.restaurant": {
+ "description": "location of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4311,
+ "in original ontology": true
+ }
+ }
+ },
+ "coffee_ordering": {
+ "description": "order a coffee drink from either Starbucks or Peets for pick up",
+ "slots": {
+ "location.store": {
+ "description": "location of the coffee store",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5510,
+ "in original ontology": true
+ },
+ "name.drink": {
+ "description": "name of the drink",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 9182,
+ "in original ontology": true
+ },
+ "size.drink": {
+ "description": "size of the drink",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 7804,
+ "in original ontology": true
+ },
+ "num.drink": {
+ "description": "number of drinks",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 848,
+ "in original ontology": true
+ },
+ "type.milk": {
+ "description": "type of the milk",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 8433,
+ "in original ontology": true
+ },
+ "preference": {
+ "description": "user preference of the drink",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 11266,
+ "in original ontology": true
+ }
+ }
+ },
+ "pizza_ordering": {
+ "description": "order a pizza",
+ "slots": {
+ "name.store": {
+ "description": "name of the pizza store",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5127,
+ "in original ontology": true
+ },
+ "name.pizza": {
+ "description": "name of the pizza",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 9208,
+ "in original ontology": true
+ },
+ "size.pizza": {
+ "description": "size of the pizza",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 9661,
+ "in original ontology": true
+ },
+ "type.topping": {
+ "description": "type of the topping",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 20639,
+ "in original ontology": true
+ },
+ "type.crust": {
+ "description": "type of the crust",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5099,
+ "in original ontology": true
+ },
+ "preference": {
+ "description": "user preference of the pizza",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4998,
+ "in original ontology": true
+ },
+ "location.store": {
+ "description": "location of the pizza store",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1487,
+ "in original ontology": true
+ }
+ }
+ },
+ "auto_repair": {
+ "description": "set up an auto repair appointment with a repair shop",
+ "slots": {
+ "name.store": {
+ "description": "name of the repair store",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4005,
+ "in original ontology": true
+ },
+ "name.customer": {
+ "description": "name of the customer",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4547,
+ "in original ontology": true
+ },
+ "date.appt": {
+ "description": "date of the appointment",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 7650,
+ "in original ontology": true
+ },
+ "time.appt": {
+ "description": "time of the appointment",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 9827,
+ "in original ontology": true
+ },
+ "reason.appt": {
+ "description": "reason of the appointment",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6509,
+ "in original ontology": true
+ },
+ "name.vehicle": {
+ "description": "name of the vehicle",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5262,
+ "in original ontology": true
+ },
+ "year.vehicle": {
+ "description": "year of the vehicle",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4561,
+ "in original ontology": true
+ },
+ "location.store": {
+ "description": "location of the repair store",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 709,
+ "in original ontology": true
+ }
+ }
+ },
+ "flights": {
+ "description": "find a round trip or multi-city flights",
+ "slots": {
+ "type": {
+ "description": "type of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1999,
+ "in original ontology": true
+ },
+ "destination1": {
+ "description": "the first destination city of the trip",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3993,
+ "in original ontology": true
+ },
+ "destination2": {
+ "description": "the second destination city of the trip",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 128,
+ "in original ontology": true
+ },
+ "origin": {
+ "description": "the origin city of the trip",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2595,
+ "in original ontology": true
+ },
+ "date.depart_origin": {
+ "description": "date of departure from origin",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3177,
+ "in original ontology": true
+ },
+ "date.depart_intermediate": {
+ "description": "date of departure from intermediate",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 48,
+ "in original ontology": true
+ },
+ "date.return": {
+ "description": "date of return",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2822,
+ "in original ontology": true
+ },
+ "time_of_day": {
+ "description": "time of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3840,
+ "in original ontology": true
+ },
+ "seating_class": {
+ "description": "seat type (first class, business class, economy class, etc.",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3626,
+ "in original ontology": true
+ },
+ "seat_location": {
+ "description": "location of the seat",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 313,
+ "in original ontology": true
+ },
+ "stops": {
+ "description": "non-stop, layovers, etc.",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6174,
+ "in original ontology": true
+ },
+ "price_range": {
+ "description": "price range of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2646,
+ "in original ontology": true
+ },
+ "num.pax": {
+ "description": "number of people",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 437,
+ "in original ontology": true
+ },
+ "luggage": {
+ "description": "luggage information",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 17,
+ "in original ontology": true
+ },
+ "total_fare": {
+ "description": "total cost of the trip",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1238,
+ "in original ontology": true
+ },
+ "other_description": {
+ "description": "other description of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2620,
+ "in original ontology": true
+ },
+ "from": {
+ "description": "departure of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1293,
+ "in original ontology": true
+ },
+ "to": {
+ "description": "destination of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1979,
+ "in original ontology": true
+ },
+ "airline": {
+ "description": "airline of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3981,
+ "in original ontology": true
+ },
+ "flight_number": {
+ "description": "the number of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 42,
+ "in original ontology": true
+ },
+ "date": {
+ "description": "date of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 756,
+ "in original ontology": true
+ },
+ "from.time": {
+ "description": "departure time of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6440,
+ "in original ontology": true
+ },
+ "to.time": {
+ "description": "arrival time of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2571,
+ "in original ontology": true
+ },
+ "stops.location": {
+ "description": "location of the stop",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1097,
+ "in original ontology": true
+ },
+ "fare": {
+ "description": "cost of the flight",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1475,
+ "in original ontology": true
+ }
+ }
+ },
+ "food_order": {
+ "description": "order take-out for a particular cuisine choice",
+ "slots": {
+ "name.item": {
+ "description": "name of the item",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6080,
+ "in original ontology": true
+ },
+ "other_description.item": {
+ "description": "other description of the item",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1474,
+ "in original ontology": true
+ },
+ "type.retrieval": {
+ "description": "type of the retrieval method",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1868,
+ "in original ontology": true
+ },
+ "total_price": {
+ "description": "total price",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5,
+ "in original ontology": true
+ },
+ "time.pickup": {
+ "description": "pick up time",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 981,
+ "in original ontology": true
+ },
+ "num.people": {
+ "description": "number of people",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 880,
+ "in original ontology": true
+ },
+ "name.restaurant": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 106,
+ "in original ontology": true
+ },
+ "type.food": {
+ "description": "type of food",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1247,
+ "in original ontology": true
+ },
+ "type.meal": {
+ "description": "type of meal",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 64,
+ "in original ontology": true
+ },
+ "location.restaurant": {
+ "description": "location of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 8,
+ "in original ontology": true
+ },
+ "rating.restaurant": {
+ "description": "rating of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5,
+ "in original ontology": true
+ },
+ "price_range": {
+ "description": "price range of the food",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5,
+ "in original ontology": true
+ }
+ }
+ },
+ "hotel": {
+ "description": "find a hotel using typical preferences",
+ "slots": {
+ "name.hotel": {
+ "description": "name of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5241,
+ "in original ontology": true
+ },
+ "location.hotel": {
+ "description": "location of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2940,
+ "in original ontology": true
+ },
+ "sub_location.hotel": {
+ "description": "rough location of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1869,
+ "in original ontology": true
+ },
+ "star_rating": {
+ "description": "star rating of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2049,
+ "in original ontology": true
+ },
+ "customer_rating": {
+ "description": "customer rating of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1239,
+ "in original ontology": true
+ },
+ "price_range": {
+ "description": "price range of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2357,
+ "in original ontology": true
+ },
+ "amenity": {
+ "description": "amenity of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 19030,
+ "in original ontology": true
+ },
+ "num.beds": {
+ "description": "number of beds to book",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 733,
+ "in original ontology": true
+ },
+ "type.bed": {
+ "description": "type of the bed",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1423,
+ "in original ontology": true
+ },
+ "num.rooms": {
+ "description": "number of rooms to book",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 256,
+ "in original ontology": true
+ },
+ "check-in_date": {
+ "description": "check-in date",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 608,
+ "in original ontology": true
+ },
+ "check-out_date": {
+ "description": "check-out date",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 428,
+ "in original ontology": true
+ },
+ "date_range": {
+ "description": "date range of the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2458,
+ "in original ontology": true
+ },
+ "num.guests": {
+ "description": "number of guests",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1323,
+ "in original ontology": true
+ },
+ "type.room": {
+ "description": "type of the room",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1840,
+ "in original ontology": true
+ },
+ "price_per_night": {
+ "description": "price per night",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2357,
+ "in original ontology": true
+ },
+ "total_fare": {
+ "description": "total fare",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 28,
+ "in original ontology": true
+ },
+ "location": {
+ "description": "location of the hotel",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 352,
+ "in original ontology": true
+ }
+ }
+ },
+ "movie": {
+ "description": "find a movie to watch in theaters or using a streaming service at home",
+ "slots": {
+ "name.movie": {
+ "description": "name of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 13413,
+ "in original ontology": true
+ },
+ "genre": {
+ "description": "genre of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4982,
+ "in original ontology": true
+ },
+ "name.theater": {
+ "description": "name of the theater",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2371,
+ "in original ontology": true
+ },
+ "location.theater": {
+ "description": "location of the theater",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2894,
+ "in original ontology": true
+ },
+ "time.start": {
+ "description": "start time of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 6455,
+ "in original ontology": true
+ },
+ "time.end": {
+ "description": "end time of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3,
+ "in original ontology": true
+ },
+ "price.ticket": {
+ "description": "price of the ticket",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 989,
+ "in original ontology": true
+ },
+ "price.streaming": {
+ "description": "price of the streaming",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 397,
+ "in original ontology": true
+ },
+ "type.screening": {
+ "description": "type of the screening",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1419,
+ "in original ontology": true
+ },
+ "audience_rating": {
+ "description": "audience rating",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1506,
+ "in original ontology": true
+ },
+ "movie_rating": {
+ "description": "film rating",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 273,
+ "in original ontology": true
+ },
+ "release_date": {
+ "description": "release date of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 386,
+ "in original ontology": true
+ },
+ "runtime": {
+ "description": "running time of the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 262,
+ "in original ontology": true
+ },
+ "real_person": {
+ "description": "name of actors, directors, etc.",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3406,
+ "in original ontology": true
+ },
+ "character": {
+ "description": "name of character in the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1676,
+ "in original ontology": true
+ },
+ "streaming_service": {
+ "description": "streaming service that provide the movie",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2729,
+ "in original ontology": true
+ },
+ "num.tickets": {
+ "description": "number of tickets",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1045,
+ "in original ontology": true
+ },
+ "seating": {
+ "description": "type of seating",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 13,
+ "in original ontology": true
+ }
+ }
+ },
+ "music": {
+ "description": "find several tracks to play and then comment on each one",
+ "slots": {
+ "name.track": {
+ "description": "name of the track",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4916,
+ "in original ontology": true
+ },
+ "name.artist": {
+ "description": "name of the artist",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 9287,
+ "in original ontology": true
+ },
+ "name.album": {
+ "description": "name of the album",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1106,
+ "in original ontology": true
+ },
+ "name.genre": {
+ "description": "music genre",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 452,
+ "in original ontology": true
+ },
+ "type.music": {
+ "description": "rough type of the music",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 603,
+ "in original ontology": true
+ },
+ "describes_track": {
+ "description": "description of a track to find",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2969,
+ "in original ontology": true
+ },
+ "describes_artist": {
+ "description": "description of a artist to find",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 612,
+ "in original ontology": true
+ },
+ "describes_album": {
+ "description": "description of an album to find",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 189,
+ "in original ontology": true
+ },
+ "describes_genre": {
+ "description": "description of a genre to find",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 26,
+ "in original ontology": true
+ },
+ "describes_type.music": {
+ "description": "description of the music type",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 375,
+ "in original ontology": true
+ }
+ }
+ },
+ "restaurant": {
+ "description": "ask for recommendations for a particular type of cuisine",
+ "slots": {
+ "name.restaurant": {
+ "description": "name of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 8676,
+ "in original ontology": true
+ },
+ "location": {
+ "description": "location of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 5165,
+ "in original ontology": true
+ },
+ "sub-location": {
+ "description": "rough location of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1210,
+ "in original ontology": true
+ },
+ "type.food": {
+ "description": "the cuisine of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 12412,
+ "in original ontology": true
+ },
+ "menu_item": {
+ "description": "item in the menu",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1499,
+ "in original ontology": true
+ },
+ "type.meal": {
+ "description": "type of meal",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2677,
+ "in original ontology": true
+ },
+ "rating": {
+ "description": "rating of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2951,
+ "in original ontology": true
+ },
+ "price_range": {
+ "description": "price range of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1930,
+ "in original ontology": true
+ },
+ "business_hours": {
+ "description": "business hours of the restaurant",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2024,
+ "in original ontology": true
+ },
+ "name.reservation": {
+ "description": "name of the person who make the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 16,
+ "in original ontology": true
+ },
+ "num.guests": {
+ "description": "number of guests",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 179,
+ "in original ontology": true
+ },
+ "time.reservation": {
+ "description": "time of the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 216,
+ "in original ontology": true
+ },
+ "date.reservation": {
+ "description": "date of the reservation",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 130,
+ "in original ontology": true
+ },
+ "type.seating": {
+ "description": "type of the seating",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 11,
+ "in original ontology": true
+ }
+ }
+ },
+ "sport": {
+ "description": "discuss facts and stats about players, teams, games, etc. in EPL, MLB, MLS, NBA, NFL",
+ "slots": {
+ "name.team": {
+ "description": "name of the team",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 19651,
+ "in original ontology": true
+ },
+ "record.team": {
+ "description": "record of the team (number of wins and losses)",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3338,
+ "in original ontology": true
+ },
+ "record.games_ahead": {
+ "description": "number of games ahead",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 33,
+ "in original ontology": true
+ },
+ "record.games_back": {
+ "description": "number of games behind",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 361,
+ "in original ontology": true
+ },
+ "place.team": {
+ "description": "ranking of the team",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4075,
+ "in original ontology": true
+ },
+ "result.match": {
+ "description": "result of the match",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3245,
+ "in original ontology": true
+ },
+ "score.match": {
+ "description": "score of the match",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 3241,
+ "in original ontology": true
+ },
+ "date.match": {
+ "description": "date of the match",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2660,
+ "in original ontology": true
+ },
+ "day.match": {
+ "description": "day of the match",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 4743,
+ "in original ontology": true
+ },
+ "time.match": {
+ "description": "time of the match",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 1283,
+ "in original ontology": true
+ },
+ "name.player": {
+ "description": "name of the player",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2365,
+ "in original ontology": true
+ },
+ "position.player": {
+ "description": "position of the player",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 2746,
+ "in original ontology": true
+ },
+ "record.player": {
+ "description": "record of the player",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 80,
+ "in original ontology": true
+ },
+ "name.non_player": {
+ "description": "name of non-palyer such as the manager, coach",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 843,
+ "in original ontology": true
+ },
+ "venue": {
+ "description": "venue of the match take place",
+ "is_categorical": false,
+ "possible_values": [],
+ "count": 328,
+ "in original ontology": true
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": ""
+ }
+ },
+ "binary_dialogue_act": [],
+ "state": {}
+}
\ No newline at end of file
diff --git a/data/unified_datasets/taskmaster/original_data.zip b/data/unified_datasets/taskmaster/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..8a36e95bf829063c7bff98404eb795107baf7b87
Binary files /dev/null and b/data/unified_datasets/taskmaster/original_data.zip differ
diff --git a/data/unified_datasets/taskmaster/preprocess.py b/data/unified_datasets/taskmaster/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..298ac2562432d5cf93dad8f6dec672e626eca249
--- /dev/null
+++ b/data/unified_datasets/taskmaster/preprocess.py
@@ -0,0 +1,445 @@
+import json
+import os
+import copy
+import zipfile
+from tqdm import tqdm
+import re
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+from pprint import pprint
+
+descriptions = {
+ "uber_lyft": {
+ "uber_lyft": "order a car for a ride inside a city",
+ "location.from": "pickup location",
+ "location.to": "destination of the ride",
+ "type.ride": "type of ride",
+ "num.people": "number of people",
+ "price.estimate": "estimated cost of the ride",
+ "duration.estimate": "estimated duration of the ride",
+ "time.pickup": "time of pickup",
+ "time.dropoff": "time of dropoff",
+ },
+ "movie_ticket": {
+ "movie_ticket": "book movie tickets for a film",
+ "name.movie": "name of the movie",
+ "name.theater": "name of the theater",
+ "num.tickets": "number of tickets",
+ "time.start": "start time of the movie",
+ "location.theater": "location of the theater",
+ "price.ticket": "price of the ticket",
+ "type.screening": "type of the screening",
+ "time.end": "end time of the movie",
+ "time.duration": "duration of the movie",
+ },
+ "restaurant_reservation": {
+ "restaurant_reservation": "searching for a restaurant and make reservation",
+ "name.restaurant": "name of the restaurant",
+ "name.reservation": "name of the person who make the reservation",
+ "num.guests": "number of guests",
+ "time.reservation": "time of the reservation",
+ "type.seating": "type of the seating",
+ "location.restaurant": "location of the restaurant",
+ },
+ "coffee_ordering": {
+ "coffee_ordering": "order a coffee drink from either Starbucks or Peets for pick up",
+ "location.store": "location of the coffee store",
+ "name.drink": "name of the drink",
+ "size.drink": "size of the drink",
+ "num.drink": "number of drinks",
+ "type.milk": "type of the milk",
+ "preference": "user preference of the drink",
+ },
+ "pizza_ordering": {
+ "pizza_ordering": "order a pizza",
+ "name.store": "name of the pizza store",
+ "name.pizza": "name of the pizza",
+ "size.pizza": "size of the pizza",
+ "type.topping": "type of the topping",
+ "type.crust": "type of the crust",
+ "preference": "user preference of the pizza",
+ "location.store": "location of the pizza store",
+ },
+ "auto_repair": {
+ "auto_repair": "set up an auto repair appointment with a repair shop",
+ "name.store": "name of the repair store",
+ "name.customer": "name of the customer",
+ "date.appt": "date of the appointment",
+ "time.appt": "time of the appointment",
+ "reason.appt": "reason of the appointment",
+ "name.vehicle": "name of the vehicle",
+ "year.vehicle": "year of the vehicle",
+ "location.store": "location of the repair store",
+ },
+ "flights": {
+ "flights": "find a round trip or multi-city flights",
+ "type": "type of the flight",
+ "destination1": "the first destination city of the trip",
+ "destination2": "the second destination city of the trip",
+ "origin": "the origin city of the trip",
+ "date.depart_origin": "date of departure from origin",
+ "date.depart_intermediate": "date of departure from intermediate",
+ "date.return": "date of return",
+ "time_of_day": "time of the flight",
+ "seating_class": "seat type (first class, business class, economy class, etc.",
+ "seat_location": "location of the seat",
+ "stops": "non-stop, layovers, etc.",
+ "price_range": "price range of the flight",
+ "num.pax": "number of people",
+ "luggage": "luggage information",
+ "total_fare": "total cost of the trip",
+ "other_description": "other description of the flight",
+ "from": "departure of the flight",
+ "to": "destination of the flight",
+ "airline": "airline of the flight",
+ "flight_number": "the number of the flight",
+ "date": "date of the flight",
+ "from.time": "departure time of the flight",
+ "to.time": "arrival time of the flight",
+ "stops.location": "location of the stop",
+ "fare": "cost of the flight",
+ },
+ "food_order": {
+ "food_order": "order take-out for a particular cuisine choice",
+ "name.item": "name of the item",
+ "other_description.item": "other description of the item",
+ "type.retrieval": "type of the retrieval method",
+ "total_price": "total price",
+ "time.pickup": "pick up time",
+ "num.people": "number of people",
+ "name.restaurant": "name of the restaurant",
+ "type.food": "type of food",
+ "type.meal": "type of meal",
+ "location.restaurant": "location of the restaurant",
+ "rating.restaurant": "rating of the restaurant",
+ "price_range": "price range of the food",
+ },
+ "hotel": {
+ "hotel": "find a hotel using typical preferences",
+ "name.hotel": "name of the hotel",
+ "location.hotel": "location of the hotel",
+ "sub_location.hotel": "rough location of the hotel",
+ "star_rating": "star rating of the hotel",
+ "customer_rating": "customer rating of the hotel",
+ "price_range": "price range of the hotel",
+ "amenity": "amenity of the hotel",
+ "num.beds": "number of beds to book",
+ "type.bed": "type of the bed",
+ "num.rooms": "number of rooms to book",
+ "check-in_date": "check-in date",
+ "check-out_date": "check-out date",
+ "date_range": "date range of the reservation",
+ "num.guests": "number of guests",
+ "type.room": "type of the room",
+ "price_per_night": "price per night",
+ "total_fare": "total fare",
+ "location": "location of the hotel",
+ },
+ "movie": {
+ "movie": "find a movie to watch in theaters or using a streaming service at home",
+ "name.movie": "name of the movie",
+ "genre": "genre of the movie",
+ "name.theater": "name of the theater",
+ "location.theater": "location of the theater",
+ "time.start": "start time of the movie",
+ "time.end": "end time of the movie",
+ "price.ticket": "price of the ticket",
+ "price.streaming": "price of the streaming",
+ "type.screening": "type of the screening",
+ "audience_rating": "audience rating",
+ "movie_rating": "film rating",
+ "release_date": "release date of the movie",
+ "runtime": "running time of the movie",
+ "real_person": "name of actors, directors, etc.",
+ "character": "name of character in the movie",
+ "streaming_service": "streaming service that provide the movie",
+ "num.tickets": "number of tickets",
+ "seating": "type of seating",
+ },
+ "music": {
+ "music": "find several tracks to play and then comment on each one",
+ "name.track": "name of the track",
+ "name.artist": "name of the artist",
+ "name.album": "name of the album",
+ "name.genre": "music genre",
+ "type.music": "rough type of the music",
+ "describes_track": "description of a track to find",
+ "describes_artist": "description of a artist to find",
+ "describes_album": "description of an album to find",
+ "describes_genre": "description of a genre to find",
+ "describes_type.music": "description of the music type",
+ },
+ "restaurant": {
+ "restaurant": "ask for recommendations for a particular type of cuisine",
+ "name.restaurant": "name of the restaurant",
+ "location": "location of the restaurant",
+ "sub-location": "rough location of the restaurant",
+ "type.food": "the cuisine of the restaurant",
+ "menu_item": "item in the menu",
+ "type.meal": "type of meal",
+ "rating": "rating of the restaurant",
+ "price_range": "price range of the restaurant",
+ "business_hours": "business hours of the restaurant",
+ "name.reservation": "name of the person who make the reservation",
+ "num.guests": "number of guests",
+ "time.reservation": "time of the reservation",
+ "date.reservation": "date of the reservation",
+ "type.seating": "type of the seating",
+ },
+ "sport": {
+ "sport": "discuss facts and stats about players, teams, games, etc. in EPL, MLB, MLS, NBA, NFL",
+ "name.team": "name of the team",
+ "record.team": "record of the team (number of wins and losses)",
+ "record.games_ahead": "number of games ahead",
+ "record.games_back": "number of games behind",
+ "place.team": "ranking of the team",
+ "result.match": "result of the match",
+ "score.match": "score of the match",
+ "date.match": "date of the match",
+ "day.match": "day of the match",
+ "time.match": "time of the match",
+ "name.player": "name of the player",
+ "position.player": "position of the player",
+ "record.player": "record of the player",
+ "name.non_player": "name of non-palyer such as the manager, coach",
+ "venue": "venue of the match take place",
+ }
+}
+
+
+def normalize_domain_name(domain):
+ if domain == 'auto':
+ return 'auto_repair'
+ elif domain == 'pizza':
+ return 'pizza_ordering'
+ elif domain == 'coffee':
+ return 'coffee_ordering'
+ elif domain == 'uber':
+ return 'uber_lyft'
+ elif domain == 'restaurant':
+ return 'restaurant_reservation'
+ elif domain == 'movie':
+ return 'movie_ticket'
+ elif domain == 'flights':
+ return 'flights'
+ elif domain == 'food-ordering':
+ return 'food_order'
+ elif domain == 'hotels':
+ return 'hotel'
+ elif domain == 'movies':
+ return 'movie'
+ elif domain == 'music':
+ return 'music'
+ elif domain == 'restaurant-search':
+ return 'restaurant'
+ elif domain == 'sports':
+ return 'sport'
+ assert 0
+
+
+def format_turns(ori_turns):
+ new_turns = []
+ previous_speaker = None
+ utt_idx = 0
+ for i, turn in enumerate(ori_turns):
+ speaker = 'system' if turn['speaker'] == 'ASSISTANT' else 'user'
+ turn['speaker'] = speaker
+ if utt_idx == 0 and speaker == 'system':
+ continue
+ if turn['text'] == '(deleted)':
+ continue
+ if not previous_speaker:
+ assert speaker != previous_speaker
+ if speaker != previous_speaker:
+ previous_speaker = speaker
+ new_turns.append(copy.deepcopy(turn))
+ utt_idx += 1
+ else:
+ # continuous speaking
+ last_turn = new_turns[-1]
+ # if ori_turns[i-1]['text'] == turn['text']:
+ # # skip repeat turn
+ # continue
+ if turn['text'] in ori_turns[i-1]['text']:
+ continue
+ index_shift = len(last_turn['text']) + 1
+ last_turn['text'] += ' '+turn['text']
+ if 'segments' in turn:
+ last_turn.setdefault('segments', [])
+ for segment in turn['segments']:
+ segment['start_index'] += index_shift
+ segment['end_index'] += index_shift
+ last_turn['segments'] += turn['segments']
+ if new_turns and new_turns[-1]['speaker'] == 'system':
+ new_turns = new_turns[:-1]
+ return new_turns
+
+
+def log_ontology(acts, ontology, ori_ontology):
+ for item in acts:
+ intent, domain, slot, value = item['intent'], item['domain'], item['slot'], item['value']
+ if domain not in ontology['domains']:
+ ontology['domains'][domain] = {'description': "", 'slots': {}}
+ if slot not in ontology['domains'][domain]['slots']:
+ ontology['domains'][domain]['slots'][slot] = {
+ 'description': '',
+ 'is_categorical': False,
+ 'possible_values': [],
+ 'count': 1
+ }
+ else:
+ ontology['domains'][domain]['slots'][slot]['count'] += 1
+ ontology['domains'][domain]['slots'][slot]['in original ontology'] = slot in ori_ontology[domain]
+ if intent is not None and intent not in ontology['intents']:
+ ontology['intents'][intent] = {
+ "description": ''
+ }
+
+
+def preprocess():
+ self_dir = os.path.dirname(os.path.abspath(__file__))
+ processed_dialogue = []
+ ontology = {'domains': {},
+ 'intents': {},
+ 'binary_dialogue_act': [],
+ 'state': {}}
+ original_zipped_path = os.path.join(self_dir, 'original_data.zip')
+ new_dir = os.path.join(self_dir, 'original_data')
+ if not os.path.exists(os.path.join(self_dir, 'data.zip')) or not os.path.exists(os.path.join(self_dir, 'ontology.json')):
+ print('unzip to', new_dir)
+ print('This may take several minutes')
+ archive = zipfile.ZipFile(original_zipped_path, 'r')
+ archive.extractall(self_dir)
+ files = [
+ ('TM-1-2019/woz-dialogs.json', 'TM-1-2019/ontology.json'),
+ ('TM-1-2019/self-dialogs.json', 'TM-1-2019/ontology.json'),
+ ('TM-2-2020/data/flights.json', 'TM-2-2020/ontology/flights.json'),
+ ('TM-2-2020/data/food-ordering.json', 'TM-2-2020/ontology/food-ordering.json'),
+ ('TM-2-2020/data/hotels.json', 'TM-2-2020/ontology/hotels.json'),
+ ('TM-2-2020/data/movies.json', 'TM-2-2020/ontology/movies.json'),
+ ('TM-2-2020/data/music.json', 'TM-2-2020/ontology/music.json'),
+ ('TM-2-2020/data/restaurant-search.json', 'TM-2-2020/ontology/restaurant-search.json'),
+ ('TM-2-2020/data/sports.json', 'TM-2-2020/ontology/sports.json')
+ ]
+ idx_count = 1
+ total = 0
+
+ for filename, ontology_filename in files:
+ data = json.load(open(os.path.join(new_dir, filename)))
+ ori_ontology = {}
+ if 'TM-1' in filename:
+ for domain, item in json.load(open(os.path.join(new_dir, ontology_filename))).items():
+ ori_ontology[item["id"]] = {}
+ for slot in item["required"] + item["optional"]:
+ ori_ontology[item["id"]][slot] = 0
+ else:
+ domain = normalize_domain_name(filename.split('/')[-1].split('.')[0])
+ ori_ontology[domain] = {}
+ for _, item in json.load(open(os.path.join(new_dir, ontology_filename))).items():
+ for group in item:
+ for anno in group["annotations"]:
+ ori_ontology[domain][anno] = 0
+ for d in ori_ontology:
+ if d not in ontology['domains']:
+ ontology['domains'][d] = {'description': descriptions[d][d], 'slots': {}}
+ for s in ori_ontology[d]:
+ if s not in ontology['domains'][d]['slots']:
+ ontology['domains'][d]['slots'][s] = {
+ 'description': descriptions[d][s],
+ 'is_categorical': False,
+ 'possible_values': [],
+ 'count': 0,
+ 'in original ontology': True
+ }
+ # pprint(ori_ontology)
+ for ori_sess in tqdm(data, desc='processing taskmaster-{}'.format(filename)):
+ total += 1
+ turns = format_turns(ori_sess['utterances'])
+ if not turns:
+ continue
+ if 'TM-2' in filename:
+ dial_domain = normalize_domain_name(filename.split('/')[-1].split('.')[0])
+ else:
+ dial_domain = normalize_domain_name(ori_sess['instruction_id'].split('-', 1)[0])
+ dialogue = {
+ "dataset": "taskmaster",
+ "data_split": "train",
+ "dialogue_id": 'taskmaster_' + str(idx_count),
+ "original_id": ori_sess['conversation_id'],
+ "instruction_id": ori_sess['instruction_id'],
+ "domains": [
+ dial_domain
+ ],
+ "turns": []
+ }
+ idx_count += 1
+ assert turns[0]['speaker'] == 'user' and turns[-1]['speaker'] == 'user', print(turns)
+ for utt_idx, uttr in enumerate(turns):
+ speaker = uttr['speaker']
+ turn = {
+ 'speaker': speaker,
+ 'utterance': uttr['text'],
+ 'utt_idx': utt_idx,
+ 'dialogue_act': {
+ 'binary': [],
+ 'categorical': [],
+ 'non-categorical': [],
+ },
+ }
+ if speaker == 'user':
+ turn['state'] = {}
+ turn['state_update'] = {'categorical': [], 'non-categorical': []}
+
+ if 'segments' in uttr:
+ for segment in uttr['segments']:
+ for item in segment['annotations']:
+ # domain = item['name'].split('.', 1)[0]
+ domain = dial_domain
+
+ # if domain != item['name'].split('.', 1)[0]:
+ # print(domain, item['name'].split('.', 1), dialogue["original_id"])
+ # assert domain in item['name'].split('.', 1)[0]
+
+ # if item['name'].split('.', 1)[0] != domain:
+ # print(domain, item['name'].split('.', 1), dialogue["original_id"])
+ slot = item['name'].split('.', 1)[-1]
+ if slot.endswith('.accept') or slot.endswith('.reject'):
+ slot = slot[:-7]
+ if slot not in ori_ontology[domain]:
+ # print(domain, item['name'].split('.', 1), dialogue["original_id"])
+ continue
+ # if domain in ori_ontology:
+ # ori_ontology[domain][slot] += 1
+ # else:
+ # print(domain, item['name'].split('.', 1), dialogue["original_id"])
+ # assert domain in ori_ontology, print(domain, item['name'].split('.', 1), dialogue["original_id"])
+
+ if not segment['text']:
+ print(slot)
+ print(segment)
+ print()
+ assert turn['utterance'][segment['start_index']:segment['end_index']] == segment['text']
+ turn['dialogue_act']['non-categorical'].append({
+ 'intent': 'inform',
+ 'domain': domain,
+ 'slot': slot,
+ 'value': segment['text'].lower(),
+ 'start': segment['start_index'],
+ 'end': segment['end_index']
+ })
+ log_ontology(turn['dialogue_act']['non-categorical'], ontology, ori_ontology)
+ dialogue['turns'].append(turn)
+ processed_dialogue.append(dialogue)
+ # pprint(ori_ontology)
+ # save ontology json
+ json.dump(ontology, open(os.path.join(self_dir, 'ontology.json'), 'w'), indent=2)
+ json.dump(processed_dialogue, open('data.json', 'w'), indent=2)
+ write_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ os.remove('data.json')
+ else:
+ # read from file
+ processed_dialogue = read_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ ontology = json.load(open(os.path.join(self_dir, 'ontology.json')))
+ return processed_dialogue, ontology
+
+if __name__ == '__main__':
+ preprocess()
diff --git a/data/unified_datasets/woz/README.md b/data/unified_datasets/woz/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..9b883e678a20fba4da78f37545acedb50945c767
--- /dev/null
+++ b/data/unified_datasets/woz/README.md
@@ -0,0 +1,31 @@
+# README
+
+## Features
+
+- Annotations: dialogue act, character-level span for non-categorical slots.
+
+Statistics:
+
+| | \# dialogues | \# utterances | avg. turns | avg. tokens | \# domains |
+| ----- | ------------ | ------------- | ---------- | ----------- | ---------- |
+| train | 406 | 2936 | 7.23 | 11.36 | 1 |
+| dev | 135 | 941 | 6.97 | 11.99 | 1 |
+| train | 135 | 935 | 6.93 | 11.87 | 1 |
+
+
+## Main changes
+
+- domain is set to **restaurant**
+- make some rule-based fixes on categorical values to make them in `possible value` lists
+- `belief_states` in WOZ dataset contains `request` intents, which are ignored in processing
+- some state annotations are not consistent with dialogue_act annotations. for example in `woz_train_en.json`, first dialog, 2nd turn:
+
+ `user: "How about Chinese food?"`
+
+ `chinese food` is included in `dialogue_act` annotation as a `inform` intent, but not updated in `belief_state` annotation.
+
+
+
+## Original data
+
+https://github.com/nmrksic/neural-belief-tracker/tree/master/data/woz
\ No newline at end of file
diff --git a/data/unified_datasets/woz/data.zip b/data/unified_datasets/woz/data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..14b3cc709f5d6c3c8361542a1ccdfef7696a436c
Binary files /dev/null and b/data/unified_datasets/woz/data.zip differ
diff --git a/data/unified_datasets/woz/ontology.json b/data/unified_datasets/woz/ontology.json
new file mode 100644
index 0000000000000000000000000000000000000000..8f863f679941e2bdf4347a4d6c992a4881b8ef60
--- /dev/null
+++ b/data/unified_datasets/woz/ontology.json
@@ -0,0 +1,117 @@
+{
+ "domains": {
+ "restaurant": {
+ "description": "search for a restaurant to dine",
+ "slots": {
+ "food": {
+ "is_categorical": false,
+ "possible_values": [],
+ "description": "food type of the restaurant"
+ },
+ "area": {
+ "is_categorical": true,
+ "possible_values": [
+ "east",
+ "west",
+ "center",
+ "north",
+ "south"
+ ],
+ "description": "area of the restaurant"
+ },
+ "postcode": {
+ "is_categorical": false,
+ "possible_values": [],
+ "description": "postal code of the restaurant"
+ },
+ "phone": {
+ "is_categorical": false,
+ "possible_values": [],
+ "description": "phone number of the restaurant"
+ },
+ "address": {
+ "is_categorical": false,
+ "possible_values": [],
+ "description": "address of the restaurant"
+ },
+ "price range": {
+ "is_categorical": true,
+ "possible_values": [
+ "expensive",
+ "moderate",
+ "cheap"
+ ],
+ "description": "price range of the restaurant"
+ },
+ "name": {
+ "is_categorical": false,
+ "possible_values": [],
+ "description": "name of the restaurant"
+ }
+ }
+ }
+ },
+ "intents": {
+ "inform": {
+ "description": "system informs user the value of a slot"
+ },
+ "request": {
+ "description": "system asks the user to provide value of a slot"
+ }
+ },
+ "binary_dialogue_act": [
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "postcode",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "phone",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "address",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "area",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "price range",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "food",
+ "value": ""
+ },
+ {
+ "intent": "request",
+ "domain": "restaurant",
+ "slot": "name",
+ "value": ""
+ }
+ ],
+ "state": {
+ "restaurant": {
+ "food": "",
+ "area": "",
+ "postcode": "",
+ "phone": "",
+ "address": "",
+ "price range": "",
+ "name": ""
+ }
+ }
+}
\ No newline at end of file
diff --git a/data/unified_datasets/woz/original_data.zip b/data/unified_datasets/woz/original_data.zip
new file mode 100644
index 0000000000000000000000000000000000000000..52f0a9c6d72e88082d0d52f64d3298516136c26a
Binary files /dev/null and b/data/unified_datasets/woz/original_data.zip differ
diff --git a/data/unified_datasets/woz/preprocess.py b/data/unified_datasets/woz/preprocess.py
new file mode 100644
index 0000000000000000000000000000000000000000..a99394b42479e927f8c2cf8194a0233bc5fe8129
--- /dev/null
+++ b/data/unified_datasets/woz/preprocess.py
@@ -0,0 +1,324 @@
+import copy
+import zipfile
+import json
+import os
+from collections import Counter
+from tqdm import tqdm
+import logging
+import sys
+sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))
+logging.basicConfig(level=logging.INFO)
+from convlab2.util.file_util import read_zipped_json, write_zipped_json
+
+self_dir = os.path.dirname(os.path.abspath(__file__))
+
+cat_slots = ['price range', 'area']
+cat_slot_values = {
+ 'area': [
+ "east",
+ "west",
+ "center",
+ "north",
+ "south"
+ ],
+ 'price range': [
+ "expensive",
+ "moderate",
+ "dontcare",
+ "cheap"
+ ]
+}
+
+woz_desc = {
+ 'restaurant': {
+ 'domain': 'search for a restaurant to dine',
+ 'food': 'food type of the restaurant',
+ 'area': 'area of the restaurant',
+ 'postcode': 'postal code of the restaurant',
+ 'phone': 'phone number of the restaurant',
+ 'address': 'address of the restaurant',
+ 'price range': 'price range of the restaurant',
+ 'name': 'name of the restaurant'
+ },
+ 'intents': {
+ 'inform': 'system informs user the value of a slot',
+ 'request': 'system asks the user to provide value of a slot',
+ }
+}
+
+
+def convert_da(da, utt, all_binary):
+ converted = {
+ 'binary': [],
+ 'categorical': [],
+ 'non-categorical': []
+ }
+
+ for s, v in da:
+ v = 'expensive' if 'expensive' in v else v
+ v = 'center' if v == 'centre' else v
+ v = 'east' if 'east' in v else v
+
+ if s in ['request']:
+ _converted = {
+ 'intent': 'request',
+ 'domain': 'restaurant',
+ 'slot': v,
+ 'value': '',
+ }
+ converted['binary'].append(_converted)
+
+ if _converted not in all_binary:
+ all_binary.append(_converted)
+
+ else:
+ slot_name = s
+ slot_type = 'categorical' if s in cat_slots else 'non-categorical'
+
+ converted[slot_type].append({
+ 'intent': 'inform',
+ 'domain': 'restaurant',
+ 'slot': slot_name,
+ 'value': v
+ })
+
+ if slot_type == 'non-categorical':
+
+ start = utt.find(v)
+
+ if start != -1:
+ end = start + len(v)
+ converted[slot_type][-1]['start'] = start
+ converted[slot_type][-1]['end'] = end
+
+ return converted
+
+
+def convert_state(state):
+ ret = {
+ 'restaurant': {}
+ }
+ for s in woz_desc['restaurant']:
+ if s == 'domain':
+ continue
+ ret['restaurant'][s] = ''
+ for s in state:
+ assert s['act'] in ['request', 'inform']
+ if s['act'] == 'inform':
+ for _s, _v in s['slots']:
+ _v = 'expensive' if 'expensive' in _v else _v
+ _v = 'center' if _v == 'centre' else _v
+ _v = 'east' if 'east' in _v else _v
+ # try:
+ # assert _s not in ret['restaurant']
+ # except:
+ # continue
+ ret['restaurant'][_s] = _v
+
+ return ret
+
+
+def get_state_update(prev_state, cur_state, usr_da, turn_idx, dialog_idx):
+
+ ret = {
+ 'categorical': [],
+ 'non-categorical': []
+ }
+ for k, v in prev_state['restaurant'].items():
+
+ if k in cur_state['restaurant'] and cur_state['restaurant'][k] == v:
+ continue
+ if k in cat_slots:
+ ret['categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': k,
+ 'value': cur_state['restaurant'][k]
+ })
+ else:
+ found = False
+ for _da in usr_da['non-categorical']:
+
+ if _da['slot'] == k and _da['value'] == cur_state['restaurant'][k]:
+ found = True
+ if v == 'dontcare':
+ ret['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': k,
+ 'value': cur_state['restaurant'][k],
+ })
+ else:
+ ret['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': k,
+ 'value': cur_state['restaurant'][k]
+ })
+
+ if 'start' in _da:
+ ret['non-categorical'][-1].update({
+ 'utt_idx': turn_idx * 2,
+ 'start': _da['start'],
+ 'end': _da['end']
+ })
+
+ if not found:
+ # print(dialog_idx, turn_idx*2)
+ # print(k, v)
+ # print('===================')
+ ret['non-categorical'].append({
+ 'domain': 'restaurant',
+ 'slot': k,
+ 'value': cur_state['restaurant'][k]
+ })
+
+ return ret
+
+
+
+def preprocess():
+ dataset_dir = 'woz'
+ data_splits = ['train', 'validate', 'test']
+ all_dialogues = []
+ all_binary_intents = []
+ all_slot = []
+ all_slot_value = {}
+ extract_dir = os.path.join(self_dir, 'original_data')
+
+ if not os.path.exists('data.zip') or not os.path.exists('ontology.json'):
+ # data not processed
+ data_zip_file = os.path.join(self_dir, 'original_data.zip')
+ if not os.path.exists(data_zip_file):
+ raise FileNotFoundError(data_zip_file)
+
+ logging.info('unzip woz data to {}'.format(extract_dir))
+ archive = zipfile.ZipFile(data_zip_file, 'r')
+ archive.extractall(extract_dir)
+
+ dialog_id = 1
+ for split in data_splits:
+
+ data = json.load(open(os.path.join(self_dir, extract_dir, 'original_data/woz_{}_en.json'.format(split))))
+
+
+ for dialogue in data:
+ ret = {}
+ ret['dataset'] = "woz"
+ ret['data_split'] = split if split != 'validate' else 'val'
+ ret['dialogue_id'] = 'woz_' + str(dialog_id)
+ ret['original_id'] = split + str(dialogue['dialogue_idx']) if split != 'validate' else 'val' + str(dialogue['dialogue_idx'])
+ ret['domains'] = ['restaurant']
+
+ ret['turns'] = []
+
+ turns = dialogue['dialogue']
+ n_turn = len(turns)
+ prev_state = {'restaurant':{k: '' for k in woz_desc['restaurant'] if k != 'domain'}}
+
+ for i in range(n_turn):
+
+ sys_utt = turns[i]['system_transcript'].lower()
+ usr_utt = turns[i]['transcript'].lower()
+ usr_da = turns[i]['turn_label']
+ bs = turns[i]['belief_state']
+
+ for s, v in usr_da:
+ if s == 'request':
+ if v not in all_slot:
+ all_slot.append(v)
+
+ if v not in all_slot_value and v != 'dontcare':
+ all_slot_value[v] = []
+
+ else:
+ if s not in all_slot:
+ all_slot.append(s)
+ if v == 'dontcare':
+ continue
+ if s not in all_slot_value:
+ all_slot_value[s] = [v]
+ else:
+ if v not in all_slot_value[s]:
+ all_slot_value[s].append(v)
+
+ if i != 0:
+ ret['turns'].append({
+ 'utt_idx': len(ret['turns']),
+ 'speaker': 'system',
+ 'utterance': sys_utt,
+ 'dialogue_act': {'binary':[], 'categorical': [], 'non-categorical':[]},
+ })
+
+ cur_state = convert_state(bs)
+ cur_usr_da = convert_da(usr_da, usr_utt, all_binary_intents)
+
+ ret['turns'].append({
+ 'utt_idx': len(ret['turns']),
+ 'speaker': 'user',
+ 'utterance': usr_utt,
+ 'state': cur_state,
+ 'dialogue_act': cur_usr_da,
+ 'state_update': get_state_update(prev_state, cur_state, cur_usr_da, i, ret['dialogue_id'])
+ })
+
+ prev_state = copy.deepcopy(cur_state)
+
+ all_dialogues.append(ret)
+ dialog_id += 1
+
+ save_file = 'data.json'
+ json.dump(all_dialogues, open(save_file, 'w'), indent=4)
+ write_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ os.remove('data.json')
+
+ new_ont = {'domains': {
+ 'restaurant': {
+ 'description': woz_desc['restaurant']['domain'],
+ 'slots': {}
+ }
+ }, 'intents': {
+ 'inform': {
+ 'description': woz_desc['intents']['inform'],
+ },
+ 'request': {
+ 'description': woz_desc['intents']['request'],
+ },
+ }, 'binary_dialogue_act': []
+ }
+ for i in all_binary_intents:
+ new_ont['binary_dialogue_act'].append(i)
+
+ for slot in all_slot_value:
+ if slot in cat_slots:
+ new_ont['domains']['restaurant']['slots'][slot] = {
+ 'is_categorical': True,
+ 'possible_values': [],
+ 'description': woz_desc['restaurant'][slot]
+ }
+ for v in all_slot_value[slot]:
+ v = 'expensive' if 'expensive' in v else v
+ v = 'center' if v == 'centre' else v
+ v = 'east' if 'east' in v else v
+ if v not in new_ont['domains']['restaurant']['slots'][slot]['possible_values']:
+ new_ont['domains']['restaurant']['slots'][slot]['possible_values'].append(v)
+ else:
+ new_ont['domains']['restaurant']['slots'][slot] = {
+ 'is_categorical': False,
+ 'possible_values': [],
+ 'description': woz_desc['restaurant'][slot]
+ }
+
+ new_ont['state'] = {
+ 'restaurant': {k: '' for k in all_slot_value}
+ }
+
+ json.dump(new_ont, open(os.path.join(self_dir, 'ontology.json'), 'w'), indent=4)
+
+ else:
+ # read from file
+ all_dialogues = read_zipped_json(os.path.join(self_dir, 'data.zip'), 'data.json')
+ new_ont = json.load(open(os.path.join(self_dir, 'ontology.json')))
+
+ return all_dialogues, new_ont
+
+
+if __name__ == '__main__':
+ preprocess()