improved langfuse tracing to every question

.gitignore

@@ -0,0 +1 @@
+.env

agent.py

@@ -13,6 +13,7 @@ from langgraph.graph.message import add_messages
 from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint
 
 from langfuse.langchain import CallbackHandler
+from langfuse.decorators import observe, langfuse_context
 from langfuse import get_client
 
 os.environ["LANGFUSE_PUBLIC_KEY"] = os.getenv("LANGFUSE_PUBLIC_KEY", "pk-lf-***")  # Public key is safe to expose in client-side code
@@ -193,19 +194,28 @@ class BasicAgent:
         print("Agent graph built successfully.")
         return agent_graph
 
-    async def __call__(self, question: str) -> str:
+    @observe(as_type="generation")
+    async def __call__(self, question: str, task_id: str = None) -> str:
         print(f"Agent received question (first 100 chars): {question[:100]}...")
-        # fixed_answer = "This is a default answer."
-        # print(f"Agent returning fixed answer: {fixed_answer}")
-        # Example query agent might receive
-        # fixed_answer = await agent.run(question)
+        
+        # Update trace context with unique identifier and metadata
+        langfuse_context.update_current_trace(
+            name=f"agent_question_{task_id or 'unknown'}",
+            user_id="agent_user",
+            session_id=task_id,
+            metadata={
+                "task_id": task_id,
+                "question_preview": question[:200]
+            },
+            tags=["agent", "question_answering"]
+        )
+        
         messages = [
             HumanMessage(
-                # content="Who is Barack Obama?"
-                # content="Divide 6790 by 5"
-                content=question # + '/nothink'
+                content=question
             )
         ]
+        
         response = await self.agent_graph.ainvoke(
             {"messages": messages}, 
             config={
@@ -213,6 +223,15 @@ class BasicAgent:
                 "callbacks": [langfuse_handler],
             }
         )
+        
         response_text = response['messages'][-1].content
-        # return response_text.split('</think>')[-1]
-        return extract_answer(response_text)
+        answer = extract_answer(response_text)
+        
+        # Update the current observation with input and output
+        langfuse_context.update_current_observation(
+            input=question,
+            output=answer,
+            metadata={"full_response_length": len(response_text)}
+        )
+        
+        return answer

app.py

@@ -93,7 +93,7 @@ async def run_and_submit_all( profile: gr.OAuthProfile | None):
             print(f"Skipping item with missing task_id or question: {item}")
             continue
         try:
-            submitted_answer = await agent(question_text)
+            submitted_answer = await agent(question_text, task_id=task_id)
             answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
         except Exception as e: