more

app.py

@@ -16,7 +16,6 @@ FastLanguageModel.for_inference(model)  # Enable optimized inference
 
 # Define the response function
 def respond(message, history, system_message, max_tokens, temperature, top_p):
-    # Extensive debugging print statements
     print("\n" + "="*50)
     print("===== RESPOND FUNCTION CALLED =====")
     print("="*50)
@@ -39,17 +38,8 @@ def respond(message, history, system_message, max_tokens, temperature, top_p):
         if history:
             print("\n--- Processing Existing History ---")
             for entry in history:
-                print(f"Processing entry: {entry}")
-                # Ensure entry is a dictionary with 'user' and 'assistant' keys
-                if isinstance(entry, dict):
-                    messages.append({"role": "user", "content": entry.get('user', '')})
-                    messages.append({"role": "assistant", "content": entry.get('assistant', '')})
-                elif isinstance(entry, list) and len(entry) == 2:
-                    # Handle case where history might be a list of [user, assistant]
-                    messages.append({"role": "user", "content": entry[0]})
-                    messages.append({"role": "assistant", "content": entry[1]})
-                else:
-                    print(f"WARNING: Unexpected history entry format: {entry}")
+                messages.append({"role": "user", "content": entry[0]})
+                messages.append({"role": "assistant", "content": entry[1]})
         
         # Add the current user message
         print("\n--- Adding Current Message ---")
@@ -95,18 +85,21 @@ def respond(message, history, system_message, max_tokens, temperature, top_p):
                 print("WARNING: System message detected in response")
                 response = "Hello! How can I assist you today?"
             
-            # Prepare return history
-            return_history = (history or []) + [
-                {"user": message, "assistant": response}
-            ]
+            # Prepare return history in OpenAI messages format
+            return_messages = []
+            for entry in (history or []):
+                return_messages.append({"role": "user", "content": entry[0]})
+                return_messages.append({"role": "assistant", "content": entry[1]})
             
-            print("\n--- Return History ---")
-            print(f"Return History Length: {len(return_history)}")
-            for entry in return_history:
-                print(f"User: {entry['user']}")
-                print(f"Assistant: {entry['assistant'][:100]}...")  # Truncate long responses
+            # Add current conversation turn
+            return_messages.append({"role": "user", "content": message})
+            return_messages.append({"role": "assistant", "content": response})
             
-            return return_history
+            print("\n--- Return Messages ---")
+            for msg in return_messages:
+                print(f"Role: {msg['role']}, Content: {msg['content'][:100]}...")
+            
+            return return_messages
         
         except Exception as gen_error:
             print("\n--- GENERATION ERROR ---")