Update functions.py

functions.py

@@ -91,38 +91,50 @@ def create_retriever_from_chroma(vectorstore_path="docs/chroma/", search_type='m
     return  retriever
 
 
-async def handle_userinput(user_question, custom_graph):
+def handle_userinput(user_question, custom_graph):
     # Add the user's question to the chat history and display it in the UI
     st.session_state.messages.append({"role": "user", "content": user_question})
     st.chat_message("user").write(user_question)
 
-    # Config setup (if required for the graph)
+    # Generate a unique thread ID for the graph's state
     config = {"configurable": {"thread_id": str(uuid.uuid4())}}
 
     try:
-        # Await the asynchronous invocation of the custom graph
-        state_dict = await custom_graph.ainvoke({"question": user_question, "steps": []}, config)
+        # Invoke the custom graph with the input question
+        state_dict =  custom_graph.invoke(
+            {"question": user_question, "steps": []}, config
+        )
 
-        # Extract documents from the state dictionary
-        docs = state_dict.get("documents", [])
+        docs = state_dict["documents"]
         with st.sidebar:
             st.subheader("Dokumentai, kuriuos Birutė gavo kaip kontekstą")
-            with st.spinner("Kraunama..."):
+            with st.spinner("Processing"):
                 for doc in docs:
-                    # Display each document
-                    st.write(f"Dokumentas: {doc}")
-
-        # Check for and display the assistant's response
-        response = state_dict.get("generation")
-        if response:
+                    # Extract document content
+                    content = doc
+                    
+                    # Extract document metadata if available
+                    #metadata =doc.metadata.get('original_doc_name', 'unknown')
+                    # Display content and metadata
+                    st.write(f"Documentas: {content}")
+                    
+                   
+        
+
+        # Check if a response (generation) was produced by the graph
+        if 'generation' in state_dict and state_dict['generation']:
+            response = state_dict["generation"]
+
+            # Add the assistant's response to the chat history and display it
             st.session_state.messages.append({"role": "assistant", "content": response})
             st.chat_message("assistant").write(response)
-        
-            
+        else:
+            st.chat_message("assistant").write("Your question violates toxicity rules or contains sensitive information.")
+
     except Exception as e:
         # Display an error message in case of failure
         st.chat_message("assistant").write("Klaida: Arba per didelis kontekstas suteiktas modeliui, arba užklausų serveryje yra per daug")
-
+        
 
 
 
@@ -691,7 +703,7 @@ def generate(state,QA_chain):
     }
 
 
-async def grade_documents(state, retrieval_grader):
+def grade_documents(state, retrieval_grader):
     question = state["question"]
     documents = state["documents"]
     steps = state["steps"]
@@ -703,7 +715,7 @@ async def grade_documents(state, retrieval_grader):
     
     for d in documents:
         # Call the grading function
-        score = await retrieval_grader.ainvoke({"question": question, "documents": d})
+        score = retrieval_grader.invoke({"question": question, "documents": d})
         print(f"Grader output for document: {score}")  # Detailed debugging output
         
         # Extract the grade