improve agenda police #31, the order of NLG could be more detailed in TemplateNLG:sorted_dialog_act

convlab2/nlg/template/multiwoz/nlg.py

@@ -2,6 +2,7 @@ import json
 import random
 import os
 from pprint import pprint
+import collections
 from convlab2.nlg import NLG
 
 
@@ -67,6 +68,32 @@ class TemplateNLG(NLG):
         self.manual_user_template = read_json(os.path.join(template_dir, 'manual_user_template_nlg.json'))
         self.manual_system_template = read_json(os.path.join(template_dir, 'manual_system_template_nlg.json'))
 
+    def sorted_dialog_act(self, dialog_acts):
+        new_action_group = {}
+        for item in dialog_acts:
+            intent, domain, slot, value = item
+            if domain not in new_action_group:
+                new_action_group[domain] = {'nooffer': [], 'inform-name': [], 'inform-other': [], 'request': [], 'other': []}
+            if intent == 'NoOffer':
+                new_action_group[domain]['nooffer'].append(item)
+            elif intent == 'Inform' and slot == 'Name':
+                new_action_group[domain]['inform-name'].append(item)
+            elif intent == 'Inform':
+                new_action_group[domain]['inform-other'].append(item)
+            elif intent == 'request':
+                new_action_group[domain]['request'].append(item)
+            else:
+                new_action_group[domain]['other'].append(item)
+
+        new_action = []
+        if 'general' in new_action_group:
+            new_action += new_action_group['general']['other']
+            del new_action_group['general']
+        for domain in new_action_group:
+            for k in ['nooffer', 'inform-name', 'inform-other', 'request', 'other']:
+                new_action = new_action_group[domain][k] + new_action
+        return new_action
+
     def generate(self, dialog_acts):
         """NLG for Multiwoz dataset
 
@@ -75,7 +102,8 @@ class TemplateNLG(NLG):
         Returns:
             generated sentence
         """
-        action = {}
+        dialog_acts = self.sorted_dialog_act(dialog_acts)
+        action = collections.OrderedDict()
         for intent, domain, slot, value in dialog_acts:
             k = '-'.join([domain, intent])
             action.setdefault(k, [])
@@ -165,9 +193,21 @@ class TemplateNLG(NLG):
                 for slot, value in slot_value_pairs:
                     if value in ["do nt care", "do n't care", "dontcare"]:
                         sentence = 'I don\'t care about the {} of the {}'.format(slot.lower(), dialog_act.split('-')[0].lower())
+                    elif self.is_user and dialog_act.split('-')[1] == 'Inform' and slot == 'Choice' and value == 'any':
+                        # user have no preference, any choice is ok
+                        sentence = random.choice([
+                            "Please pick one for me. ",
+                            "Anyone would be ok. ",
+                            "Just select one for me. "
+                        ])
                     elif dialog_act in template and slot in template[dialog_act]:
                         sentence = random.choice(template[dialog_act][slot])
                         sentence = sentence.replace('#{}-{}#'.format(dialog_act.upper(), slot.upper()), str(value))
+                    elif slot == 'NotBook':
+                        sentence = random.choice([
+                            "I do not need to book. ",
+                            "I 'm not looking to make a booking at the moment."
+                        ])
                     else:
                         if slot in slot2word:
                             sentence = 'The {} is {} . '.format(slot2word[slot], str(value))
@@ -201,7 +241,7 @@ class TemplateNLG(NLG):
 
 def example():
     # dialog act
-    dialog_acts = [['Inform', 'Train', 'Day', 'wednesday'], ['Inform', 'Train', 'Leave', '10:15']]
+    dialog_acts = [['Inform', 'Hotel', 'Area', 'east'],['Inform', 'Hotel', 'Name', 'fds'], ['welcome', 'general', 'none', 'none']]
     print(dialog_acts)
 
     # system model for manual, auto, auto_manual

convlab2/policy/rule/multiwoz/policy_agenda_multiwoz.py

@@ -74,11 +74,11 @@ class UserPolicyAgendaMultiWoz(Policy):
         self.domain_goals = self.goal.domain_goals
         self.agenda = Agenda(self.goal)
 
-    def predict(self, state):
+    def predict(self, sys_dialog_act):
         """
         Predict an user act based on state and preorder system action.
         Args:
-            state (tuple): Dialog state.
+            sys_dialog_act (list): system dialogue act: [[intent, domain, slot, value],...].
         Returns:
             action (tuple): User act.
             session_over (boolean): True to terminate session, otherwise session continues.
@@ -86,10 +86,10 @@ class UserPolicyAgendaMultiWoz(Policy):
         """
         self.__turn += 2
 
-        assert isinstance(state, list)
+        assert isinstance(sys_dialog_act, list)
 
         sys_action = {}
-        for intent, domain, slot, value in state:
+        for intent, domain, slot, value in sys_dialog_act:
             k = '-'.join([domain, intent])
             sys_action.setdefault(k,[])
             sys_action[k].append([slot, value])
@@ -143,29 +143,35 @@ class UserPolicyAgendaMultiWoz(Policy):
 
     @classmethod
     def _transform_usract_out(cls, action):
+        # print('before transform', action)
         new_action = {}
         for act in action.keys():
             if '-' in act:
                 if 'general' not in act:
                     (dom, intent) = act.split('-')
                     new_act = dom.capitalize() + '-' + intent.capitalize()
-                    if action[act] == [['none', 'none']]:
-                        new_action[new_act] = [['none', 'none']]
-                        continue
                     new_action[new_act] = []
                     for pairs in action[act]:
                         slot = REF_USR_DA_M[dom.capitalize()].get(pairs[0], None)
-                        if slot is not None:
+                        if pairs[0] == 'none' and pairs[1] == 'none':
+                            new_action[new_act].append(['none', 'none'])
+                        elif pairs[0] == 'choice' and pairs[1] == 'any':
+                            new_action[new_act].append(['Choice', 'any'])
+                        elif pairs[0] == 'NotBook' and pairs[1] == 'none':
+                            new_action[new_act].append(['NotBook', 'none'])
+                        elif slot is not None:
                             new_action[new_act].append([slot, pairs[1]])
                     # new_action[new_act] = [[REF_USR_DA_M[dom.capitalize()].get(pairs[0], pairs[0]), pairs[1]] for pairs in action[act]]
                 else:
                     new_action[act] = action[act]
             else:
                 pass
+        # print('after transform', new_action)
         return new_action
 
     @classmethod
     def _transform_sysact_in(cls, action):
+        # print("sys in", action)
         new_action = {}
         if not isinstance(action, dict):
             logging.warning('illegal da: {}'.format(action))
@@ -195,7 +201,7 @@ class UserPolicyAgendaMultiWoz(Policy):
                         new_action[act.lower()] = new_list
             else:
                 new_action[act.lower()] = action[act]
-
+        # print("sys in transformed", new_action)
         return new_action
 
     @classmethod
@@ -209,6 +215,9 @@ class UserPolicyAgendaMultiWoz(Policy):
         if slot not in cls.stand_value_dict[domain]:
             return value
 
+        if slot in ['parking', 'internet'] and value == 'none':
+            return 'yes'
+
         value_list = cls.stand_value_dict[domain][slot]
         low_value_list = [item.lower() for item in value_list]
         value_list = sorted(list(set(value_list)|set(low_value_list)))
@@ -386,6 +395,8 @@ class Agenda(object):
                                           len(goal.domain_goals[domain]['info'])):
                     self.__push(domain + '-inform', slot, goal.domain_goals[domain]['info'][slot])
 
+            self.__push(domain + '-inform', "none", "none")
+
         self.cur_domain = None
 
     def update(self, sys_action, goal: Goal):
@@ -422,6 +433,21 @@ class Agenda(object):
                 if self.update_domain(diaact, slot_vals, goal):
                     return
 
+        for diaact in sys_action.keys():
+            if 'inform' in diaact or 'recommend' in diaact:
+                for slot, val in sys_action[diaact]:
+                    if slot == 'name':
+                        self._remove_item(diaact.split('-')[0]+'-inform', 'choice')
+            if 'booking' in diaact:
+                g_book = self._get_goal_infos(self.cur_domain, goal)[-2]
+                if len(g_book) == 0:
+                    self._push_item(self.cur_domain + '-inform', "NotBook", "none")
+            if 'OfferBook' in diaact:
+                domain = diaact.split('-')[0]
+                g_book = self._get_goal_infos(domain, goal)[-2]
+                if len(g_book) == 0:
+                    self._push_item(domain + '-inform', "NotBook", "none")
+
         self.post_process(goal)
 
     def post_process(self, goal: Goal):
@@ -517,6 +543,7 @@ class Agenda(object):
         Returns:
             action (dict): user diaact
         """
+        # print(self)
         diaacts, slots, values = self.__pop(initiative)
         action = {}
         for (diaact, slot, value) in zip(diaacts, slots, values):
@@ -563,6 +590,10 @@ class Agenda(object):
                     logging.warning('illegal booking slot: {}, domain: {}'.format(slot, domain))
                     continue
 
+            # For multiple choices, add new intent to select one:
+            if slot == 'choice':
+                self._push_item(domain + '-inform', "choice", "any")
+
             if slot in g_reqt:
                 if not self._check_reqt_info(domain):
                     self._remove_item(domain + '-request', slot)
@@ -628,10 +659,13 @@ class Agenda(object):
                             goal.domain_goals[places[-2]]['reqt']['address'] not in NOT_SURE_VALS:
                         self._push_item(domain + '-inform', slot, goal.domain_goals[places[-2]]['reqt']['address'])
 
-                    elif random.random() < 0.5:
-                        self._push_item(domain + '-inform', slot, DEF_VAL_DNC)
+                    # elif random.random() < 0.5:
+                    #     self._push_item(domain + '-inform', slot, DEF_VAL_DNC)
 
-                elif random.random() < 0.5:
+                # elif random.random() < 0.5:
+                #     self._push_item(domain + '-inform', slot, DEF_VAL_DNC)
+
+                # for those sys requests that are not in user goal
                 self._push_item(domain + '-inform', slot, DEF_VAL_DNC)
         return False
 
@@ -668,6 +702,9 @@ class Agenda(object):
     def _handle_select(self, domain, intent, slot_vals, goal: Goal):
         g_reqt, g_info, g_fail_info, g_book, g_fail_book = self._get_goal_infos(domain, goal)
         # delete Choice
+        for slot, val in slot_vals:
+            if slot == 'choice':
+                self._push_item(domain + '-inform', "choice", "any")
         slot_vals = [[slot, val] for [slot, val] in slot_vals if slot != 'choice']
 
         if slot_vals:
@@ -814,3 +851,83 @@ class Agenda(object):
         text += '<stack btm>\n'
         text += '-----agenda-----\n'
         return text
+
+
+if __name__ == '__main__':
+    import numpy as np
+    import torch
+    np.random.seed(42)
+    random.seed(42)
+    torch.manual_seed(42)
+
+    user_policy = UserPolicyAgendaMultiWoz()
+    from convlab2.policy.rule.multiwoz.rule_based_multiwoz_bot import RuleBasedMultiwozBot
+    sys_policy = RuleBasedMultiwozBot()
+    from convlab2.nlg.template.multiwoz.nlg import TemplateNLG
+    user_nlg = TemplateNLG(is_user=True, mode='manual')
+    sys_nlg = TemplateNLG(is_user=False, mode='manual')
+    from convlab2.util.multiwoz.state import default_state
+
+    user_policy.init_session()
+    sys_policy.init_session()
+
+    print(user_policy.goal)
+
+    print(user_policy.agenda)
+    user_act = user_policy.predict([])
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    state = default_state()
+    state['user_action'] = user_act
+    sys_act = sys_policy.predict(state)
+    sys_act.append(["Request", "Restaurant", "Price", "?"])
+    print(sys_act)
+
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    print(sys_act)
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    sys_act = [["Book", "Booking", "Ref", "7GAWK763"]]
+    print(sys_act)
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    sys_act = [["Reqmore", "General", "none", "none"]]
+    print(sys_act)
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    sys_act = [["Inform", "Hotel", "Parking", "none"]]
+    print(sys_act)
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    sys_act = [["Request", "Booking", "people", "?"]]
+    print(sys_act)
+
+    user_act = user_policy.predict(sys_act)
+    print(user_act)
+    user_utt = user_nlg.generate(user_act)
+    print(user_utt)
+    sys_act = sys_policy.predict(state)
+    sys_act = [["Inform", "Hotel", "Post", "233"], ["Book", "Booking", "none", "none"]]
+    print(sys_act)