From b82732eae951b3dc957136f40b992a1904c9cbe5 Mon Sep 17 00:00:00 2001
From: zhuqi <zqwerty@users.noreply.github.com>
Date: Wed, 2 Sep 2020 15:45:16 +0800
Subject: [PATCH] Add police n hospital (#95)
* add back police and hospital goal
* update police db:add postcode; update hospital db:add address and postcode; update dbquery: query hospital with department, deepcopy query result
---
convlab2/task/multiwoz/goal_generator.py | 24 +--
convlab2/util/multiwoz/dbquery.py | 19 +-
data/multiwoz/db/hospital_db.json | 264 +++++++++++++++++------
data/multiwoz/db/police_db.json | 3 +-
4 files changed, 224 insertions(+), 86 deletions(-)
diff --git a/convlab2/task/multiwoz/goal_generator.py b/convlab2/task/multiwoz/goal_generator.py
index e543818..cdf378b 100755
--- a/convlab2/task/multiwoz/goal_generator.py
+++ b/convlab2/task/multiwoz/goal_generator.py
@@ -135,14 +135,14 @@ class GoalGenerator:
"""User goal generator."""
def __init__(self,
- goal_model_path=os.path.join(get_root_path(), 'data/multiwoz/goal/new_goal_model_no_police_hospital.pkl'),
+ goal_model_path=os.path.join(get_root_path(), 'data/multiwoz/goal/new_goal_model.pkl'),
corpus_path=None,
boldify=False,
sample_info_from_trainset=True,
sample_reqt_from_trainset=False):
"""
Args:
- goal_model_path: path to a goal model
+ goal_model_path: path to a goal model
corpus_path: path to a dialog corpus to build a goal model
boldify: highlight some information in the goal message
sample_info_from_trainset: if True, sample info slots combination from train set, else sample each slot independently
@@ -163,13 +163,13 @@ class GoalGenerator:
self._build_goal_model()
print('Building goal model is done')
- # remove some slot (now no police and hospital domains)
- # del self.ind_slot_dist['police']['reqt']['postcode']
- # del self.ind_slot_value_dist['police']['reqt']['postcode']
- # del self.ind_slot_dist['hospital']['reqt']['postcode']
- # del self.ind_slot_value_dist['hospital']['reqt']['postcode']
- # del self.ind_slot_dist['hospital']['reqt']['address']
- # del self.ind_slot_value_dist['hospital']['reqt']['address']
+ # remove some slot
+ del self.ind_slot_dist['police']['reqt']['postcode']
+ del self.ind_slot_value_dist['police']['reqt']['postcode']
+ del self.ind_slot_dist['hospital']['reqt']['postcode']
+ del self.ind_slot_value_dist['hospital']['reqt']['postcode']
+ del self.ind_slot_dist['hospital']['reqt']['address']
+ del self.ind_slot_value_dist['hospital']['reqt']['address']
# print(self.slots_combination_dist['police'])
# print(self.slots_combination_dist['hospital'])
@@ -187,8 +187,6 @@ class GoalGenerator:
domain_orderings = []
for d in dialogs:
d_domains = _get_dialog_domains(dialogs[d])
- if 'police' in d_domains or 'hospital' in d_domains:
- continue
first_index = []
for domain in d_domains:
message = [dialogs[d]['goal']['message']] if type(dialogs[d]['goal']['message']) == str else \
@@ -211,9 +209,6 @@ class GoalGenerator:
self.slots_num_dist = {domain: {} for domain in domains}
for d in dialogs:
- d_domains = _get_dialog_domains(dialogs[d])
- if 'police' in d_domains or 'hospital' in d_domains:
- continue
for domain in domains:
if dialogs[d]['goal'][domain] != {}:
domain_cnt[domain] += 1
@@ -740,5 +735,4 @@ class GoalGenerator:
if __name__ == '__main__':
goal_generator = GoalGenerator(corpus_path=os.path.join(get_root_path(), 'data/multiwoz/train.json'), sample_reqt_from_trainset=True)
- # goal_generator._build_goal_model()
pprint(goal_generator.get_user_goal())
diff --git a/convlab2/util/multiwoz/dbquery.py b/convlab2/util/multiwoz/dbquery.py
index c3a5c2f..b84bf8a 100755
--- a/convlab2/util/multiwoz/dbquery.py
+++ b/convlab2/util/multiwoz/dbquery.py
@@ -3,6 +3,7 @@
import json
import os
import random
+from copy import deepcopy
class Database(object):
@@ -26,9 +27,16 @@ class Database(object):
'taxi_types': random.choice(self.dbs[domain]['taxi_types']),
'taxi_phone': ''.join([str(random.randint(1, 9)) for _ in range(11)])}]
if domain == 'police':
- return self.dbs['police']
+ return deepcopy(self.dbs['police'])
if domain == 'hospital':
- return self.dbs['hospital']
+ department = None
+ for key, val in constraints:
+ if key == 'department':
+ department = val
+ if not department:
+ return deepcopy(self.dbs['hospital'])
+ else:
+ return [deepcopy(x) for x in self.dbs['hospital'] if x['department'].lower() == department.strip().lower()]
for ele in constraints:
if ele[0] == 'area' and ele[1] == 'center':
ele[1] = 'centre'
@@ -62,10 +70,13 @@ class Database(object):
except:
continue
else:
- record['Ref'] = '{0:08d}'.format(i)
- found.append(record)
+ res = deepcopy(record)
+ res['Ref'] = '{0:08d}'.format(i)
+ found.append(res)
return found
+
+
if __name__ == '__main__':
db = Database()
print(db.query("train", [['departure', 'cambridge'], ['destination','peterborough'], ['day', 'tuesday'], ['arriveBy', '11:15']]))
diff --git a/data/multiwoz/db/hospital_db.json b/data/multiwoz/db/hospital_db.json
index 82fe62e..f8873e3 100755
--- a/data/multiwoz/db/hospital_db.json
+++ b/data/multiwoz/db/hospital_db.json
@@ -2,331 +2,463 @@
{
"department": "neurosciences critical care unit",
"id": 0,
- "phone": "01223216297"
+ "phone": "01223216297",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "trauma high dependency unit",
"id": 1,
- "phone": "01223216305"
+ "phone": "01223216305",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "neurology neurosurgery",
"id": 2,
- "phone": "01223217216"
+ "phone": "01223217216",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "oncology neurosurgery",
"id": 3,
- "phone": "01223216314"
+ "phone": "01223216314",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "children's oncology and haematology",
"id": 4,
- "phone": "01223217231"
+ "phone": "01223217231",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "children's surgical and medicine",
"id": 5,
- "phone": "01223217450"
+ "phone": "01223217450",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "acute medicine for the elderly",
"id": 6,
- "phone": "01223217261"
+ "phone": "01223217261",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "general medicine and nephrology",
"id": 7,
- "phone": "01223217195"
+ "phone": "01223217195",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medicine for the elderly",
"id": 8,
- "phone": "01223216988"
+ "phone": "01223216988",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "hepatobillary and gastrointestinal surgery regional referral centre",
"id": 9,
- "phone": "01223217300"
+ "phone": "01223217300",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "trauma and orthopaedics",
"id": 10,
- "phone": "01223217279"
+ "phone": "01223217279",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "teenage cancer trust unit",
"id": 11,
- "phone": "01223274222"
+ "phone": "01223274222",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "haematology and haematological oncology",
"id": 12,
- "phone": "01223217312"
+ "phone": "01223217312",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "children's surgical and medicine",
"id": 13,
- "phone": "01223217250"
+ "phone": "01223217250",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "intermediate dependancy area",
"id": 14,
- "phone": "01223348144"
+ "phone": "01223348144",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "hepatology",
"id": 15,
- "phone": "01223217712"
+ "phone": "01223217712",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "haematology",
"id": 16,
- "phone": "01223274679"
+ "phone": "01223274679",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "neurology",
"id": 17,
- "phone": "01223274680"
+ "phone": "01223274680",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "surgery",
"id": 18,
- "phone": "01223217303"
+ "phone": "01223217303",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "trauma and orthopaedics",
"id": 19,
- "phone": "01223217282"
+ "phone": "01223217282",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "oncology",
"id": 20,
- "phone": "01223217708"
+ "phone": "01223217708",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "infectious diseases",
"id": 21,
- "phone": "01223217314"
+ "phone": "01223217314",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "haematology day unit",
"id": 22,
- "phone": "01223348169"
+ "phone": "01223348169",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "clinical decisions unit",
"id": 23,
- "phone": "01223596203"
+ "phone": "01223596203",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "acute medical assessment unit",
"id": 24,
- "phone": "01223348314"
+ "phone": "01223348314",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medical short stay unit",
"id": 25,
- "phone": "01223348336"
+ "phone": "01223348336",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "inpatient occupational therapy",
"id": 26,
- "phone": "01223216881"
+ "phone": "01223216881",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "paediatric day unit",
"id": 27,
- "phone": "01223217567"
+ "phone": "01223217567",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "paediatric clinic",
"id": 28,
- "phone": "01223348313"
+ "phone": "01223348313",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medicine for the elderly",
"id": 29,
- "phone": "01223274744"
+ "phone": "01223274744",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "transplant high dependency unit",
"id": 30,
- "phone": "01223216811"
+ "phone": "01223216811",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "diabetes and endocrinology",
"id": 31,
- "phone": "01223217323"
+ "phone": "01223217323",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "infusion services",
"id": 32,
- "phone": "01223586967"
+ "phone": "01223586967",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medicine for the elderly",
"id": 33,
- "phone": "01223217484"
+ "phone": "01223217484",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medicine for the elderly",
"id": 34,
- "phone": "01223217498"
+ "phone": "01223217498",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "transplant unit",
"id": 35,
- "phone": "01223217711"
+ "phone": "01223217711",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medicine for the elderly",
"id": 36,
- "phone": "01223217428"
+ "phone": "01223217428",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "theatre admissions unit",
"id": 37,
- "phone": "01223256583"
+ "phone": "01223256583",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "cardiology",
"id": 38,
- "phone": "01223256233"
+ "phone": "01223256233",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "major trauma unit",
"id": 39,
- "phone": "01223349881"
+ "phone": "01223349881",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "cardiology and coronary care unit",
"id": 40,
- "phone": "01223256459"
+ "phone": "01223256459",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "colorectal surgery",
"id": 41,
- "phone": "01223348545"
+ "phone": "01223348545",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "plastic and vascular surgery plastics",
"id": 42,
- "phone": "01223274280"
+ "phone": "01223274280",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "gastroenterology",
"id": 43,
- "phone": "01223274284"
+ "phone": "01223274284",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "oral and maxillofacial surgery and ent",
"id": 44,
- "phone": "01223348527"
+ "phone": "01223348527",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "urology",
"id": 45,
- "phone": "01223256650"
+ "phone": "01223256650",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "respiratory medicine",
"id": 46,
- "phone": "01223256645"
+ "phone": "01223256645",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "lewin stroke and rehabilitation unit",
"id": 47,
- "phone": "01223274750"
+ "phone": "01223274750",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "neurosciences",
"id": 48,
- "phone": "01223216348"
+ "phone": "01223216348",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "psychiatry",
"id": 49,
- "phone": "01223596201"
+ "phone": "01223596201",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "emergency department",
"id": 50,
- "phone": "01223217118"
+ "phone": "01223217118",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "cambridge eye unit",
"id": 51,
- "phone": "01223257168"
+ "phone": "01223257168",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "clinical investigation ward",
"id": 52,
- "phone": "01223586706"
+ "phone": "01223586706",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "clinical research facility",
"id": 53,
- "phone": "01223596055"
+ "phone": "01223596055",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "coronary care unit",
"id": 54,
- "phone": "01223217297"
+ "phone": "01223217297",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "intermediate dependency area",
"id": 55,
- "phone": "01223217873"
+ "phone": "01223217873",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "medical decisions unit",
"id": 56,
- "phone": "01223596066"
+ "phone": "01223596066",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "paediatric intensive care unit",
"id": 57,
- "phone": "01223217715"
+ "phone": "01223217715",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "paediatric day unit",
"id": 58,
- "phone": "01223257157"
+ "phone": "01223257157",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "john farman intensive care unit",
"id": 59,
- "phone": "01223256166"
+ "phone": "01223256166",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "delivery unit",
"id": 60,
- "phone": "01223217217"
+ "phone": "01223217217",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "postnatal",
"id": 61,
- "phone": "01223217667"
+ "phone": "01223217667",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "neonatal unit",
"id": 62,
- "phone": "01223217678"
+ "phone": "01223217678",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "antenatal",
"id": 63,
- "phone": "01223217671"
+ "phone": "01223217671",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "transitional care",
"id": 64,
- "phone": "01223254668"
+ "phone": "01223254668",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
},
{
"department": "gynaecology",
"id": 65,
- "phone": "01223257206"
+ "phone": "01223257206",
+ "address": "hills rd , cambridge",
+ "postcode": "cb20qq"
}
]
diff --git a/data/multiwoz/db/police_db.json b/data/multiwoz/db/police_db.json
index dbe208f..e040cec 100755
--- a/data/multiwoz/db/police_db.json
+++ b/data/multiwoz/db/police_db.json
@@ -3,6 +3,7 @@
"name": "Parkside Police Station",
"address": "Parkside, Cambridge",
"id": 0,
- "phone": "01223358966"
+ "phone": "01223358966",
+ "postcode": "cb11jg"
}
]
--
GitLab