chatbot/main.py

@@ -1,5 +1,6 @@
 from pathlib import Path
-from modules.nlp import NaturalLanguageProcessor
+from modules.nlu import NLU, Slot, UserAct
+from modules.state_monitor import DialogStateMonitor
 from modules.generator import ResponseGenerator
 from modules.config import Config
 import colorama
@@ -13,7 +14,8 @@ def main():
     config_path = base_path / 'config' / 'config.json'
     config = Config.load_config(config_path)
 
-    nlp = NaturalLanguageProcessor(config)
+    nlu = NLU()
+    dst = DialogStateMonitor()
     generator = ResponseGenerator(config)
 
     print(Fore.CYAN + "Witaj w chatbocie! Rozpocznij rozmowę.")
@@ -25,10 +27,14 @@ def main():
             print(Fore.RED + "Zamykanie chatbota...")
             break
 
-        intent = nlp.analyze(user_input)
-        response = generator.generate(intent)
+        user_act = nlu.analyze(user_input)
+        # user_act = UserAct(intent='inform',
+        #                    slots=[Slot(name='item', value='laptop'), Slot(name='item', value='kot'),Slot(name='address', value='123 Main St')])
+        dst.update(user_act)
+        print(dst.state)
+        # response = generator.generate(intent)
 
-        print(Fore.CYAN + "Bot: " + response)
+        # print(Fore.CYAN + "Bot: " + response)
 
 
 if __name__ == "__main__":

chatbot/modules/nlu.py

@@ -1,15 +1,39 @@
 from flair.models import SequenceTagger
 import sys
+
 sys.path.append("..")
 from models.nlu_train2 import predict_frame, predict_slot
 import logging
 
 logging.getLogger('flair').setLevel(logging.CRITICAL)
 
+
+class Slot:
+    def __init__(self, name, value=None):
+        self.name = name
+        self.value = value
+
+    def __str__(self) -> str:
+        return f"Name: {self.name}, Value: {self.value}"
+
+
+class UserAct:
+    def __init__(self, intent: str, slots: list[Slot] = []):
+        self.slots = slots
+        self.intent = intent
+
+    def __str__(self):
+        msg = f"Act: {self.intent}, Slots: ["
+        for slot in self.slots:
+            msg += f"({slot}), "
+        msg += "]"
+        return msg
+
+
 class NLU:
     def __init__(self):
-        self.frame_model = SequenceTagger.load('../models/frame-model/final-model.pt')
-        self.slot_model = SequenceTagger.load('../models/slot-model/final-model.pt')
+        self.frame_model = SequenceTagger.load('models/frame-model/final-model.pt')
+        self.slot_model = SequenceTagger.load('models/slot-model/final-model.pt')
 
     def get_intent(self, text: str):
         return predict_frame(self.frame_model, text.split(), 'frame')
@@ -24,14 +48,14 @@ class NLU:
             slot = frame["slot"]
             if slot.startswith("B-"):
                 if current_slot:
-                    slots.append({'name': current_slot, 'value': " ".join(current_slot_value)})
+                    slots.append(Slot(name=current_slot, value=current_slot_value))
                 current_slot = slot[2:]
                 current_slot_value = [frame["form"]]
             elif slot.startswith("I-"):
                 current_slot_value.append(frame["form"])
 
         if current_slot:
-            slots.append({'name': current_slot, 'value': " ".join(current_slot_value)})
+            slots.append(Slot(name=current_slot, value=current_slot_value))
 
         return slots
 
@@ -40,11 +64,4 @@ class NLU:
         slots = self.get_slot(text)
         print({'intent': intent,
                'slots': slots})
-        return {
-            'intent': intent,
-            'slots': slots
-        }
-
-nlu = NLU()
-
-nlu.analyze("Chce kupic lakier do pazanokci")
+        return UserAct(intent=intent, slots=slots)

chatbot/modules/state_monitor.py

@@ -1,6 +1,66 @@
-class DialogueStateMonitor:
-    def __init__(self) -> None:
-        self.state = {'last_intent': 'unknown'}
+import copy
+from modules.nlu import UserAct
+import json
 
-    def update_state(self, intent: str) -> None:
-        self.state['last_intent'] = intent
+
+class DialogStateMonitor:
+    def __init__(self):
+        self.__initial_state = dict(
+            belief_state={
+                'item': {},
+                'address': {},
+                'card_nr': {},
+                'delivery_method': {},
+                'payment_method': {},
+                'email': {},
+                'order-complete': False,
+            },
+            act='',
+            slot_names=[])
+
+        self.state = copy.deepcopy(self.__initial_state)
+
+    def is_value_empty(self, d, key):
+        value = d.get(key, None)
+        if value in [None, '', [], {}]:
+            return True
+        return False
+
+    def update_act(self, intent):
+        self.state['act'] = intent
+
+    def update_slot_names(self, slots_names):
+        self.state['slot_names'] = slots_names
+
+    def check_order_complete(self):
+        all_filled = all(bool(self.state['belief_state'][key]) for key in
+                         ['item', 'address', 'card_nr', 'delivery_method', 'payment_method', 'email'])
+        self.state['belief_state']['order-complete'] = all_filled
+
+    def update(self, act: UserAct) -> None:
+        print(act)
+        if act.intent == 'inform':
+            self.update_act(act.intent)
+            slots_mapping = {
+                'item': [],
+                'address': [],
+                'card_nr': [],
+                'delivery_method': [],
+                'payment_method': [],
+                'email': []
+            }
+            for slot in act.slots:
+                if slot.name in slots_mapping and self.is_value_empty(self.state, slot.name):
+                    slots_mapping[slot.name].append(slot.value)  # To do: normalization
+
+            for slot_name, values in slots_mapping.items():
+                if values:
+                    self.state['belief_state'][slot_name] = values
+        elif act.intent == 'request':
+            self.update_act(act.intent)
+            slots_names = [slot.name for slot in act.slots]
+            self.update_slot_names(slots_names)
+        elif act.intent == 'bye':
+            self.update_act(act.intent)
+
+        self.check_order_complete()