update

app/services/chat_processor.py

@@ -1,7 +1,6 @@
 from datetime import datetime, timezone
 from typing import Optional, Dict, Any
-from concurrent.futures import ThreadPoolExecutor, as_completed
-from functools import lru_cache
+from concurrent.futures import ThreadPoolExecutor
 from yake import KeywordExtractor
 from app.services.chathistory import ChatSession
 from app.services.websearch import WebSearch
@@ -10,10 +9,7 @@ from app.services.environmental_condition import EnvironmentalData
 from app.services.prompts import *
 from app.services.vector_database_search import VectorDatabaseSearch
 import re
-
-# Create a singleton instance outside the class
 vectordb = VectorDatabaseSearch()
-model = Model()  # Reuse model instance
 
 class ChatProcessor:
     def __init__(self, token: str, session_id: Optional[str] = None, num_results: int = 3, num_images: int = 3):
@@ -27,24 +23,18 @@ class ChatProcessor:
         self.environment_data = EnvironmentalData(city)
         self.web_searcher = WebSearch(num_results=num_results, max_images=num_images)
         self.web_search_required = True
-        self._keyword_extractors = {}  # Cache keyword extractors
 
-    @lru_cache(maxsize=128)
-    def get_keyword_extractor(self, language: str, max_ngram_size: int, num_keywords: int):
-        """Cache keyword extractors for reuse"""
+    def extract_keywords_yake(self, text: str, language: str, max_ngram_size: int = 2, num_keywords: int = 4) -> list:
         lang_code = "en"
         if language.lower() == "urdu":
             lang_code = "ur"
-            
-        return KeywordExtractor(
+
+        kw_extractor = KeywordExtractor(
             lan=lang_code,
             n=max_ngram_size,
             top=num_keywords,
             features=None
         )
-
-    def extract_keywords_yake(self, text: str, language: str, max_ngram_size: int = 2, num_keywords: int = 4) -> list:
-        kw_extractor = self.get_keyword_extractor(language, max_ngram_size, num_keywords)
         keywords = kw_extractor.extract_keywords(text)
         return [kw[0] for kw in keywords]
 
@@ -64,48 +54,36 @@ class ChatProcessor:
 
     def process_chat(self, query: str) -> Dict[str, Any]:
         try:
-            # Fetch profile and history in parallel
-            with ThreadPoolExecutor(max_workers=2) as executor:
-                future_profile = executor.submit(self.chat_session.get_name_and_age)
-                future_history = executor.submit(self.chat_session.load_chat_history)
-                
-                profile = future_profile.result()
-                future_history.result()  # Wait for history to load
-                
+            profile = self.chat_session.get_name_and_age()
             name = profile['name']
             age = profile['age']
-            
-            # Update title and get history format
-            self.chat_session.update_title(self.session_id, query)
+            self.chat_session.load_chat_history()
+            self.chat_session.update_title(self.session_id,query)
             history = self.chat_session.format_history()
 
-            # Get enhanced query 
-            history_based_prompt = HISTORY_BASED_PROMPT.format(history=history, query=query)
-            enhanced_query = model.send_message_openrouter(history_based_prompt)
+            history_based_prompt = HISTORY_BASED_PROMPT.format(history=history,query= query)
+
+            enhanced_query = Model().send_message_openrouter(history_based_prompt)
 
-            # Ensure session and get user preferences
             self.session_id = self.ensure_valid_session(title=enhanced_query)
             permission = self.chat_session.get_user_preferences()
-            
-            # Extract permissions
-            websearch_enabled = permission.get('websearch', False)
+            websearch_enabled  = permission.get('websearch', False)
             env_recommendations = permission.get('environmental_recommendations', False)
             personalized_recommendations = permission.get('personalized_recommendations', False)
             keywords_permission = permission.get('keywords', False)
             reference_permission = permission.get('references', False)
             language = self.chat_session.get_language().lower()
 
-            language_prompt = LANGUAGE_RESPONSE_PROMPT.format(language=language)
 
-            if websearch_enabled:
-                # Run web and image search in parallel
+            language_prompt = LANGUAGE_RESPONSE_PROMPT.format(language = language)
+
+            if websearch_enabled :
                 with ThreadPoolExecutor(max_workers=2) as executor:
                     future_web = executor.submit(self.web_searcher.search, enhanced_query)
                     future_images = executor.submit(self.web_searcher.search_images, enhanced_query)
                     web_results = future_web.result()
                     image_results = future_images.result()
 
-                # Efficiently build context and references
                 context_parts = []
                 references = []
 
@@ -116,27 +94,13 @@ class ChatProcessor:
 
                 context = "\n".join(context_parts)
 
-                # Pre-fetch user details and environmental data in parallel if needed
-                if env_recommendations or personalized_recommendations:
-                    with ThreadPoolExecutor(max_workers=2) as executor:
-                        futures = {}
-                        if personalized_recommendations:
-                            futures['user_details'] = executor.submit(self.chat_session.get_personalized_recommendation)
-                        if env_recommendations:
-                            futures['env_data'] = executor.submit(self.environment_data.get_environmental_data)
-                        
-                        # Get results as they complete
-                        user_details = futures.get('user_details').result() if 'user_details' in futures else None
-                        env_data = futures.get('env_data').result() if 'env_data' in futures else None
-
-                # Select appropriate prompt
                 if env_recommendations and personalized_recommendations:
                     prompt = ENVIRONMENTAL_PERSONALIZED_PROMPT.format(
                         user_name=name,
                         user_age=age,
                         history=history,
-                        user_details=user_details,
-                        environmental_condition=env_data,
+                        user_details=self.chat_session.get_personalized_recommendation(),
+                        environmental_condition=self.environment_data.get_environmental_data(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
@@ -145,16 +109,16 @@ class ChatProcessor:
                     prompt = PERSONALIZED_PROMPT.format(
                         user_name=name,
                         user_age=age,
-                        user_details=user_details,
+                        user_details=self.chat_session.get_personalized_recommendation(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
                     )
-                elif env_recommendations:
+                elif env_recommendations :
                     prompt = ENVIRONMENTAL_PROMPT.format(
                         user_name=name,
                         user_age=age,
-                        environmental_condition=env_data,
+                        environmental_condition=self.environment_data.get_environmental_data(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
@@ -166,12 +130,16 @@ class ChatProcessor:
                         current_query=enhanced_query
                     )
 
-                prompt += language_prompt
-                response = model.llm(prompt, enhanced_query)
+                prompt = prompt + language_prompt
 
-                # Process keywords only if needed
-                keywords = self.extract_keywords_yake(response, language=language) if keywords_permission else ""
-                references = "" if not reference_permission else references
+                response = Model().llm(prompt,enhanced_query)
+
+                keywords = ""
+
+                if (keywords_permission):
+                    keywords = self.extract_keywords_yake(response, language=language)
+                if (not reference_permission):
+                    references = ""
 
                 chat_data = {
                     "query": enhanced_query,
@@ -190,56 +158,42 @@ class ChatProcessor:
                 return chat_data
             
             else:
-                # Start RAG timing
-                start_time = datetime.now(timezone.utc)
-                
-                # Launch image search in the background while doing vector search
+                attach_image = False
+
                 with ThreadPoolExecutor(max_workers=2) as executor:
                     future_images = executor.submit(self.web_searcher.search_images, enhanced_query)
-                    future_vector = executor.submit(vectordb.search, enhanced_query, 3)
-                    
-                    results = future_vector.result()
                     image_results = future_images.result()
 
-                # Process vector search results
+                start_time = datetime.now(timezone.utc)
+
+                results = vectordb.search( query=enhanced_query, top_k=3)
+
                 context_parts = []
                 references = []
-                seen_pages = set()
-                attach_image = False
+                seen_pages = set() 
 
                 for result in results:
                     confidence = result['confidence']
                     if confidence > 60:
                         context_parts.append(f"Content: {result['content']}")
                         page = result['page']
-                        if page not in seen_pages:
+                        if page not in seen_pages:  # Only append if page is not seen
                             references.append(f"Source: {result['source']}, Page: {page}")
                             seen_pages.add(page)
-                        attach_image = True
-
-                context = "\n".join(context_parts) if context_parts else "There is no context found unfortunately"
-
-                # Pre-fetch user details and environmental data in parallel if needed
-                if env_recommendations or personalized_recommendations:
-                    with ThreadPoolExecutor(max_workers=2) as executor:
-                        futures = {}
-                        if personalized_recommendations:
-                            futures['user_details'] = executor.submit(self.chat_session.get_personalized_recommendation)
-                        if env_recommendations:
-                            futures['env_data'] = executor.submit(self.environment_data.get_environmental_data)
-                        
-                        # Get results as they complete
-                        user_details = futures.get('user_details').result() if 'user_details' in futures else None
-                        env_data = futures.get('env_data').result() if 'env_data' in futures else None
-
-                # Select appropriate prompt
-                if env_recommendations and personalized_recommendations:
+                        attach_image = True 
+
+                context = "\n".join(context_parts)
+
+                if not context or len(context) < 10:
+                    context = "There is no context found unfortunately"
+
+                if env_recommendations  and personalized_recommendations:
                     prompt = ENVIRONMENTAL_PERSONALIZED_PROMPT.format(
                         user_name=name,
-                        user_age=age,
+                        user_age = age,
                         history=history,
-                        user_details=user_details,
-                        environmental_condition=env_data,
+                        user_details=self.chat_session.get_personalized_recommendation(),
+                        environmental_condition=self.environment_data.get_environmental_data(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
@@ -248,16 +202,16 @@ class ChatProcessor:
                     prompt = PERSONALIZED_PROMPT.format(
                         user_name=name,
                         user_age=age,
-                        user_details=user_details,
+                        user_details=self.chat_session.get_personalized_recommendation(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
                     )
-                elif env_recommendations:
+                elif env_recommendations :
                     prompt = ENVIRONMENTAL_PROMPT.format(
                         user_name=name,
                         user_age=age,
-                        environmental_condition=env_data,
+                        environmental_condition=self.environment_data.get_environmental_data(),
                         previous_history=history,
                         context=context,
                         current_query=enhanced_query
@@ -269,23 +223,24 @@ class ChatProcessor:
                         current_query=enhanced_query
                     )
 
-                prompt += language_prompt
-                response = model.llm(prompt, query)
+                prompt = prompt + language_prompt
+
+                response = Model().response = Model().llm(prompt,query)
+
                 end_time = datetime.now(timezone.utc)
 
-                # Process keywords only if needed
                 keywords = ""
-                if keywords_permission and attach_image:
+
+                if (keywords_permission):
                     keywords = self.extract_keywords_yake(response, language=language)
                     
-                references = "" if not reference_permission else references
-                image_results = "" if not attach_image else image_results
+                if (not reference_permission):
+                    references = ""
 
-                # Find where to truncate the response if needed
-                match = re.search(r'(## Personal Recommendations|## Environmental Considerations)', response)
-                truncated_response = response[:match.start()].strip() if match else response
+                if  not attach_image:
+                    image_results = ""
+                    keywords = ""
 
-                # Prepare chat data
                 chat_data = {
                     "query": enhanced_query,
                     "response": response,
@@ -297,26 +252,15 @@ class ChatProcessor:
                     "timestamp": datetime.now(timezone.utc).isoformat(),
                     "session_id": self.chat_session.session_id
                 }
-
-                # Save in parallel
-                with ThreadPoolExecutor(max_workers=2) as executor:
-                    future_save_details = executor.submit(
-                        self.chat_session.save_details,
-                        session_id=self.session_id,
-                        context=context,
-                        query=enhanced_query,
-                        response=truncated_response,
-                        rag_start_time=start_time,
-                        rag_end_time=end_time
-                    )
-                    future_save_chat = executor.submit(self.chat_session.save_chat, chat_data)
-                    
-                    # Check results
-                    if not future_save_details.result():
-                        raise ValueError("Failed to save the RAG details")
-                    if not future_save_chat.result():
-                        raise ValueError("Failed to save chat message")
-                
+                match = re.search(r'(## Personal Recommendations|## Environmental Considerations)', response)
+                if match:
+                    truncated_response = response[:match.start()].strip()
+                else:
+                    truncated_response = response
+                if not self.chat_session.save_details(session_id=self.session_id , context= context , query= enhanced_query , response=truncated_response , rag_start_time=start_time  , rag_end_time=end_time ):
+                    raise ValueError("Failed to save the RAG details")
+                if not self.chat_session.save_chat(chat_data):
+                    raise ValueError("Failed to save chat message")
                 return chat_data
 
         except Exception as e:

pyproject.toml

@@ -41,7 +41,8 @@ dependencies = [
     "python-pptx==1.0.2",
     "puremagic==1.28",
     "charset-normalizer==3.4.1",
-    "pytesseract==0.3.13"
+    "pytesseract==0.3.13",
+    "functools"
 ]
 
 [build-system]