Refactor classifier functions to include conversation context and imp…

…rove topic extraction handling

apps/ai_api/eda_ai_api/api/routes/classifier.py

@@ -31,58 +31,58 @@
 )
 
 
-def extract_topics(message: str) -> list[str]:
-    """Extract topics from message"""
-    response = llm.complete(TOPIC_TEMPLATE.format(message=message))
+async def extract_topics(message: str, history: list) -> list[str]:
+    """Extract topics with conversation context"""
+    context = "\n".join([f"{msg['role']}: {msg['content']}" for msg in history])
+    response = llm.complete(TOPIC_TEMPLATE.format(context=context, message=message))
+    if response.text.strip() == "INSUFFICIENT_CONTEXT":
+        return ["INSUFFICIENT_CONTEXT"]
     topics = [t.strip() for t in response.text.split(",") if t.strip()][:5]
-    return topics if topics else ["AI", "Technology"]
+    return topics if topics else ["INSUFFICIENT_CONTEXT"]
 
 
-def extract_proposal_details(message: str) -> tuple[str, str]:
-    """Extract project and grant details"""
-    response = llm.complete(PROPOSAL_TEMPLATE.format(message=message))
+async def extract_proposal_details(message: str, history: list) -> tuple[str, str]:
+    """Extract project and grant details with conversation context"""
+    context = "\n".join([f"{msg['role']}: {msg['content']}" for msg in history])
+    response = llm.complete(PROPOSAL_TEMPLATE.format(context=context, message=message))
     extracted = response.text.split("|")
     community_project = extracted[0].strip() if len(extracted) > 0 else "unknown"
     grant_call = extracted[1].strip() if len(extracted) > 1 else "unknown"
     return community_project, grant_call
 
 
-def process_decision(decision: str, message: str) -> Dict[str, Any]:
-    """Process routing decision and return result"""
+async def process_decision(
+    decision: str, message: str, history: list
+) -> Dict[str, Any]:
+    """Process routing decision with conversation context"""
     logger.info(f"Processing decision: {decision} for message: {message}")
 
     if decision == "discovery":
-        topics = extract_topics(message)
+        topics = await extract_topics(message, history)
         logger.info(f"Extracted topics: {topics}")
 
-        if topics == ["AI", "Technology"]:
+        if topics == ["INSUFFICIENT_CONTEXT"]:
+            context = "\n".join([f"{msg['role']}: {msg['content']}" for msg in history])
             response = llm.complete(
-                INSUFFICIENT_TEMPLATES["discovery"].format(message=message)
+                INSUFFICIENT_TEMPLATES["discovery"].format(
+                    context=context, message=message
+                )
             )
             return {"response": response.text}
 
-        return (
-            OpportunityFinderCrew().crew().kickoff(inputs={"topics": ", ".join(topics)})
-        )
-
     elif decision == "proposal":
-        community_project, grant_call = extract_proposal_details(message)
+        community_project, grant_call = await extract_proposal_details(message, history)
         logger.info(f"Project: {community_project}, Grant: {grant_call}")
 
         if community_project == "unknown" or grant_call == "unknown":
+            context = "\n".join([f"{msg['role']}: {msg['content']}" for msg in history])
             response = llm.complete(
-                INSUFFICIENT_TEMPLATES["proposal"].format(message=message)
+                INSUFFICIENT_TEMPLATES["proposal"].format(
+                    context=context, message=message
+                )
             )
             return {"response": response.text}
 
-        return (
-            ProposalWriterCrew(
-                community_project=community_project, grant_call=grant_call
-            )
-            .crew()
-            .kickoff()
-        )
-
     elif decision == "heartbeat":
         return {"is_alive": True}
 
@@ -140,7 +140,7 @@ async def classifier_route(
         decision = response.text.strip().lower()
 
         # Process decision and store conversation
-        result = process_decision(decision, combined_message)
+        result = await process_decision(decision, combined_message, history)
         await zep.add_conversation(session_id, combined_message, str(result))
 
         return ClassifierResponse(result=str(result), session_id=session_id)

apps/ai_api/eda_ai_api/utils/prompts.py

@@ -14,46 +14,59 @@
 )
 
 TOPIC_TEMPLATE = PromptTemplate(
-    """Extract up to 5 most relevant topics for grant opportunity research from the user message.
-    Return only a comma-separated list of topics (maximum 5), no other text.
+    """Previous conversation:
+{context}
 
-    User message: {message}
+Extract up to 5 most relevant topics for grant opportunity research from the user message.
+If there isn't enough context to determine specific topics, return "INSUFFICIENT_CONTEXT".
+Otherwise, return only a comma-separated list of topics (maximum 5), no other text.
+
+User message: {message}
 
-    Topics:"""
+Output:"""
 )
 
 PROPOSAL_TEMPLATE = PromptTemplate(
-    """Extract the community project name and grant program name from the user message.
-    Return in format: project_name|grant_name
-    If either cannot be determined, use "unknown" as placeholder.
+    """Previous conversation:
+{context}
 
-    User message: {message}
+Extract the community project name and grant program name from the user message.
+Return in format: project_name|grant_name
+If either cannot be determined, use "unknown" as placeholder.
 
-    Output:"""
+User message: {message}
+
+Output:"""
 )
 
 INSUFFICIENT_TEMPLATES = {
     "proposal": PromptTemplate(
-        """The user wants to write a proposal but hasn't provided enough information.
-        Generate a friendly response in the same language as the user's message asking for:
-        1. The project name and brief description
-        2. The specific grant program they're applying to (if any)
-        3. The main objectives of their project
-        4. The target community or region
-        
-        User message: {message}
-        
-        Response:"""
+        """Previous conversation:
+{context}
+
+The user wants to write a proposal but hasn't provided enough information.
+Generate a friendly response in the same language as the user's message asking for:
+1. The project name and brief description
+2. The specific grant program they're applying to (if any)
+3. The main objectives of their project
+4. The target community or region
+
+User message: {message}
+
+Response:"""
     ),
     "discovery": PromptTemplate(
-        """The user wants to find grant opportunities but hasn't provided enough information.
-        Generate a friendly response in the same language as the user's message asking for:
-        1. The main topics or areas of their project
-        2. The target region or community
-        3. Any specific funding requirements or preferences
-        
-        User message: {message}
-        
-        Response:"""
+        """Previous conversation:
+{context}
+
+The user wants to find grant opportunities but hasn't provided enough information.
+Generate a friendly response in the same language as the user's message asking for:
+1. The main topics or areas of their project
+2. The target region or community
+3. Any specific funding requirements or preferences
+
+User message: {message}
+
+Response:"""
     ),
 }