fix

ai_service/api.py

@@ -7,6 +7,8 @@ from fastapi import FastAPI, HTTPException
 from fastapi.middleware.cors import CORSMiddleware
 from pydantic import BaseModel, Field
 from typing import List, Optional
+import uuid
+from datetime import datetime, timezone
 
 from ai_service.models.model_service import model_service
 from ai_service.models.chat_service import chat_service
@@ -120,7 +122,7 @@ async def test_ollama_connection():
 
     try:
         # Try to connect to Ollama API
-        response = requests.get(f"{config.OLLAMA_API_URL}/api/tags", timeout=5)
+        response = requests.get(f"{config.OLLAMA_API_URL}/api/tags", timeout=config.API_TIMEOUT)
         response.raise_for_status()
 
         # Return the models from Ollama
@@ -130,6 +132,106 @@ async def test_ollama_connection():
             "ollama_url": config.OLLAMA_API_URL,
             "models": response.json()
         }
+    except requests.exceptions.Timeout as e:
+        return {
+            "status": "error",
+            "message": f"Timeout connecting to Ollama API: {str(e)}. The request exceeded the {config.API_TIMEOUT} second timeout.",
+            "ollama_url": config.OLLAMA_API_URL
+        }
+    except requests.exceptions.ConnectionError as e:
+        return {
+            "status": "error",
+            "message": f"Connection error to Ollama API: {str(e)}. Please check if Ollama is running at {config.OLLAMA_API_URL}.",
+            "ollama_url": config.OLLAMA_API_URL
+        }
+    except Exception as e:
+        return {
+            "status": "error",
+            "message": f"Failed to connect to Ollama API: {str(e)}",
+            "ollama_url": config.OLLAMA_API_URL
+        }
+
+@app.post("/test-chat")
+async def test_chat_completion():
+    """
+    Test the chat completion with a simple prompt.
+
+    Returns:
+        Model response.
+    """
+    try:
+        # Use the model service directly
+        response = model_service.generate_response(
+            model_id=config.DEFAULT_MODEL,
+            prompt="Hello, how are you?",
+            context=[],
+            use_rag=False
+        )
+
+        return {
+            "status": "success",
+            "model": config.DEFAULT_MODEL,
+            "response": response,
+            "ollama_url": config.OLLAMA_API_URL
+        }
+    except Exception as e:
+        return {
+            "status": "error",
+            "message": f"Failed to get chat completion: {str(e)}",
+            "ollama_url": config.OLLAMA_API_URL
+        }
+
+@app.post("/test-ollama-direct")
+async def test_ollama_direct():
+    """
+    Test the Ollama API directly with a simple chat request.
+
+    Returns:
+        Raw Ollama API response.
+    """
+    import requests
+
+    try:
+        # Prepare a simple chat request
+        request_json = {
+            "model": config.DEFAULT_MODEL,
+            "messages": [
+                {"role": "system", "content": "You are a helpful assistant."},
+                {"role": "user", "content": "Hello, how are you?"}
+            ],
+            "stream": False
+        }
+
+        # Make the API call to Ollama
+        print(f"Sending direct request to Ollama API at: {config.OLLAMA_API_URL}/api/chat")
+        response = requests.post(
+            f"{config.OLLAMA_API_URL}/api/chat",
+            headers={"Content-Type": "application/json"},
+            json=request_json,
+            timeout=config.API_TIMEOUT
+        )
+
+        response.raise_for_status()
+        result = response.json()
+
+        return {
+            "status": "success",
+            "ollama_url": config.OLLAMA_API_URL,
+            "request": request_json,
+            "response": result
+        }
+    except requests.exceptions.Timeout as e:
+        return {
+            "status": "error",
+            "message": f"Timeout connecting to Ollama API: {str(e)}. The request exceeded the {config.API_TIMEOUT} second timeout.",
+            "ollama_url": config.OLLAMA_API_URL
+        }
+    except requests.exceptions.ConnectionError as e:
+        return {
+            "status": "error",
+            "message": f"Connection error to Ollama API: {str(e)}. Please check if Ollama is running at {config.OLLAMA_API_URL}.",
+            "ollama_url": config.OLLAMA_API_URL
+        }
     except Exception as e:
         return {
             "status": "error",
@@ -151,6 +253,7 @@ async def get_config():
         "openwebui_url": config.OPENWEBUI_URL,
         "ollama_api_url": config.OLLAMA_API_URL,
         "default_model": config.DEFAULT_MODEL,
+        "api_timeout": config.API_TIMEOUT,
         "available_models": list(model_service.AVAILABLE_MODELS.keys())
     }
 
@@ -251,6 +354,11 @@ async def send_message(chat_id: str, request: MessageRequest):
         Bot response message.
     """
     try:
+        print(f"Processing message for chat {chat_id} from user {request.user_id}")
+        print(f"Message: {request.message[:50]}...")  # Print first 50 chars of message
+        print(f"Using RAG: {request.use_rag}")
+        print(f"Model parameters: temperature={request.temperature}, max_tokens={request.max_tokens}")
+
         # Extract model parameters from the request
         response = chat_service.get_chat_response(
             chat_id=chat_id,
@@ -270,10 +378,25 @@ async def send_message(chat_id: str, request: MessageRequest):
             function_calling=request.function_calling
         )
 
+        print(f"Response received. Length: {len(response.get('content', ''))}")
         return response
 
     except ValueError as e:
-        raise HTTPException(status_code=404, detail=str(e))
+        error_msg = f"Chat not found: {str(e)}"
+        print(f"ERROR: {error_msg}")
+        raise HTTPException(status_code=404, detail=error_msg)
+    except Exception as e:
+        error_msg = f"Error processing message: {str(e)}"
+        print(f"ERROR: {error_msg}")
+        # Return an error message instead of raising an exception
+        # This ensures the client gets a proper response
+        return {
+            "id": str(uuid.uuid4()),
+            "content": f"Error processing message: {str(e)}",
+            "user_id": None,
+            "is_user_message": False,
+            "timestamp": datetime.now(timezone.utc).isoformat()
+        }
 
 @app.post("/chats/{chat_id}/members/{user_id}")
 async def add_team_member(chat_id: str, user_id: str):

ai_service/config.py

@@ -24,6 +24,7 @@ class Config:
     # Ollama configuration
     OLLAMA_API_URL = os.environ.get('OLLAMA_API_URL', 'http://127.0.0.1:11434')
     DEFAULT_MODEL = os.environ.get('DEFAULT_MODEL', 'llama3.1')
+    API_TIMEOUT = int(os.environ.get('API_TIMEOUT', 300))  # Default timeout of 5 minutes (300 seconds)
 
     # Document processing
     CHUNK_SIZE = int(os.environ.get('CHUNK_SIZE', 1000))

ai_service/models/chat_service.py

@@ -286,21 +286,35 @@ class ChatService:
 
         # Get response from model
         model_id = chat['model_id']
-        response_text = model_service.generate_response(
+        try:
-            model_id=model_id,
+            print(f"Requesting response from model {model_id} for chat {chat_id}")
-            prompt=message,
+            response_text = model_service.generate_response(
-            context=context,
+                model_id=model_id,
-            use_rag=use_rag,
+                prompt=message,
-            model_params=model_params
+                context=context,
-        )
+                use_rag=use_rag,
+                model_params=model_params
+            )
 
-        # Add bot response to chat
+            # Add bot response to chat
-        response_message = self.add_message(
+            response_message = self.add_message(
-            chat_id=chat_id,
+                chat_id=chat_id,
-            content=response_text,
+                content=response_text,
-            user_id=user_id,
+                user_id=user_id,
-            is_user_message=False
+                is_user_message=False
-        )
+            )
+
+        except Exception as e:
+            error_message = f"Error generating response: {str(e)}"
+            print(f"ERROR: {error_message}")
+
+            # Add error message to chat
+            response_message = self.add_message(
+                chat_id=chat_id,
+                content=error_message,
+                user_id=user_id,
+                is_user_message=False
+            )
 
         return response_message
 

ai_service/models/model_service.py

@@ -52,6 +52,7 @@ class ModelService:
         self.ollama_api_url = config.OLLAMA_API_URL
         self.openwebui_url = config.OPENWEBUI_URL
         self.openwebui_api_key = config.OPENWEBUI_API_KEY
+        self.api_timeout = config.API_TIMEOUT
 
     def get_available_models(self) -> List[Dict[str, Any]]:
         """
@@ -193,7 +194,7 @@ class ModelService:
                     f"{self.openwebui_url}/api/chat/completions",
                     headers=headers,
                     json=openwebui_request,
-                    timeout=60  # Longer timeout for RAG
+                    timeout=self.api_timeout
                 )
 
                 response.raise_for_status()
@@ -205,9 +206,19 @@ class ModelService:
                 else:
                     return "Error: Unexpected response format from OpenWebUI"
 
+            except requests.exceptions.Timeout as e:
+                error_msg = f"Timeout error connecting to OpenWebUI API: {str(e)}. The request exceeded the {self.api_timeout} second timeout."
+                print(f"ERROR: {error_msg}")
+                print("Falling back to direct Ollama call without RAG")
+                # Continue to the Ollama API call below
+            except requests.exceptions.ConnectionError as e:
+                error_msg = f"Connection error to OpenWebUI API: {str(e)}. Please check if OpenWebUI is running at {self.openwebui_url}."
+                print(f"ERROR: {error_msg}")
+                print("Falling back to direct Ollama call without RAG")
+                # Continue to the Ollama API call below
             except Exception as e:
-                print(f"Error calling OpenWebUI API: {str(e)}")
+                error_msg = f"Error calling OpenWebUI API: {str(e)}"
-                # Fall back to direct Ollama call without RAG
+                print(f"ERROR: {error_msg}")
                 print("Falling back to direct Ollama call without RAG")
                 # Continue to the Ollama API call below
 
@@ -250,7 +261,7 @@ class ModelService:
                 f"{self.ollama_api_url}/api/chat",
                 headers={"Content-Type": "application/json"},
                 json=request_json,
-                timeout=60
+                timeout=self.api_timeout
             )
 
             response.raise_for_status()
@@ -263,8 +274,18 @@ class ModelService:
             else:
                 return "Error: Unexpected response format from Ollama"
 
+        except requests.exceptions.Timeout as e:
+            error_msg = f"Timeout error connecting to Ollama API: {str(e)}. The request exceeded the {self.api_timeout} second timeout."
+            print(f"ERROR: {error_msg}")
+            return error_msg
+        except requests.exceptions.ConnectionError as e:
+            error_msg = f"Connection error to Ollama API: {str(e)}. Please check if Ollama is running at {self.ollama_api_url}."
+            print(f"ERROR: {error_msg}")
+            return error_msg
         except Exception as e:
-            return f"Error generating response: {str(e)}"
+            error_msg = f"Error generating response: {str(e)}"
+            print(f"ERROR: {error_msg}")
+            return error_msg
 
 
 # Create a singleton instance