Fix issues

ai_service/openwebui_api.py

@@ -23,7 +23,7 @@ async def get_models():
     Get available models in OpenWebUI-compatible format.
     """
     models = model_service.get_available_models()
-    
+
     # Convert to OpenWebUI format
     openwebui_models = []
     for model in models:
@@ -36,7 +36,23 @@ async def get_models():
             "root": model["id"],
             "parent": None
         })
-    
+
+    # Debug log
+    print(f"OpenWebUI models: {json.dumps(openwebui_models, indent=2)}")
+
+    # Ensure we're returning a properly formatted list
+    if not openwebui_models:
+        # Return at least one default model if none are found
+        return [{
+            "id": "llama3.1",
+            "object": "model",
+            "created": int(time.time()),
+            "owned_by": "user",
+            "permission": [],
+            "root": "llama3.1",
+            "parent": None
+        }]
+
     return openwebui_models
 
 # Chat completions endpoint (OpenAI-compatible)
@@ -47,7 +63,7 @@ async def chat_completions(request: Request):
     """
     # Parse request body
     body = await request.json()
-    
+
     # Extract parameters
     model_id = body.get("model", "llama3.1")
     messages = body.get("messages", [])
@@ -58,26 +74,26 @@ async def chat_completions(request: Request):
     frequency_penalty = body.get("frequency_penalty")
     presence_penalty = body.get("presence_penalty")
     stop = body.get("stop")
-    
+
     # Create a unique chat ID
     chat_id = str(uuid.uuid4())
-    
+
     # Create a user ID (in a real implementation, this would come from authentication)
     user_id = "openwebui-user"
-    
+
     # Create a new chat
     chat_service.create_chat(user_id=user_id, title="API Chat", model_id=model_id)
-    
+
     # Extract the user's message (last user message in the array)
     user_message = None
     for msg in reversed(messages):
         if msg.get("role") == "user":
             user_message = msg.get("content")
             break
-    
+
     if not user_message:
         raise HTTPException(status_code=400, detail="No user message found")
-    
+
     # Get chat response
     response = chat_service.get_chat_response(
         chat_id=chat_id,
@@ -90,10 +106,10 @@ async def chat_completions(request: Request):
         presence_penalty=presence_penalty,
         stop_sequences=stop if isinstance(stop, list) else [stop] if stop else None
     )
-    
+
     # Format response in OpenAI-compatible format
     completion_id = f"chatcmpl-{str(uuid.uuid4())[:8]}"
-    
+
     openai_response = {
         "id": completion_id,
         "object": "chat.completion",
@@ -115,16 +131,16 @@ async def chat_completions(request: Request):
             "total_tokens": -1
         }
     }
-    
+
     # Handle streaming if requested
     if stream:
         async def generate_stream():
             # Yield the response in the SSE format
             yield f"data: {json.dumps(openai_response)}\n\n"
             yield "data: [DONE]\n\n"
-        
+
         return StreamingResponse(generate_stream(), media_type="text/event-stream")
-    
+
     return openai_response
 
 # Health check endpoint
@@ -143,27 +159,27 @@ async def ollama_generate(request: Request):
     """
     # Parse request body
     body = await request.json()
-    
+
     # Extract parameters
     model_id = body.get("model", "llama3.1")
     prompt = body.get("prompt", "")
-    
+
     # Create a unique chat ID
     chat_id = str(uuid.uuid4())
-    
+
     # Create a user ID (in a real implementation, this would come from authentication)
     user_id = "openwebui-user"
-    
+
     # Create a new chat
     chat_service.create_chat(user_id=user_id, title="API Chat", model_id=model_id)
-    
+
     # Get chat response
     response = chat_service.get_chat_response(
         chat_id=chat_id,
         message=prompt,
         user_id=user_id
     )
-    
+
     # Format response in Ollama-compatible format
     ollama_response = {
         "model": model_id,
@@ -171,5 +187,5 @@ async def ollama_generate(request: Request):
         "response": response.get("content", ""),
         "done": True
     }
-    
+
     return ollama_response