Update app.py

app.py

@@ -3,10 +3,10 @@ from pydantic import BaseModel
 from llama_cpp import Llama
 from concurrent.futures import ThreadPoolExecutor, as_completed
 import uvicorn
-from dotenv import load_dotenv
-import re
 import huggingface_hub
 import spaces
+import re
+from dotenv import load_dotenv
 
 load_dotenv()
 
@@ -56,9 +56,8 @@ class ModelManager:
     def load_model(self, model_config):
         try:
             return {"model": Llama.from_pretrained(repo_id=model_config['repo_id'], filename=model_config['filename']), "name": model_config['name']}
-        except Exception as e:
-            print(f"Error loading model {model_config['name']}: {e}")
-            return None
+        except Exception:
+            pass
     
     def load_all_models(self):
         if self.loaded:
@@ -77,7 +76,6 @@ class ModelManager:
         return self.models
 
 model_manager = ModelManager()
-global_data['models'] = model_manager.load_all_models()
 
 class ChatRequest(BaseModel):
     message: str
@@ -97,8 +95,8 @@ def generate_chat_response(request, model_data):
         )
         reply = response['choices'][0]['message']['content']
         return {"response": reply, "literal": user_input, "model_name": model_data['name']}
-    except Exception as e:
-        return {"response": f"Error: {str(e)}", "literal": user_input, "model_name": model_data['name']}
+    except Exception:
+        pass
 
 def normalize_input(input_text):
     return input_text.strip()
@@ -129,27 +127,29 @@ def select_best_response(responses):
 
 @app.post("/generate")
 def generate_chat(request: ChatRequest):
-    responses = []
-    num_models = len(global_data['models'])
-    
-    with ThreadPoolExecutor() as executor:
-        futures = [executor.submit(generate_chat_response, request, model_data) for model_data in global_data['models']]
-        for future in as_completed(futures):
-            try:
-                response = future.result()
-                responses.append(response)
-            except Exception as e:
-                print(f"Error in response generation: {e}")
-    
-    if not responses:
-        raise HTTPException(status_code=500, detail="Error: No responses generated.")
-    
-    best_response = select_best_response(responses)
-    
-    return {
-        "best_response": best_response,
-        "all_responses": responses
-    }
+    try:
+        global_data['models'] = model_manager.load_all_models()
+        responses = []
+        with ThreadPoolExecutor() as executor:
+            futures = [executor.submit(generate_chat_response, request, model_data) for model_data in global_data['models']]
+            for future in as_completed(futures):
+                try:
+                    response = future.result()
+                    if response:
+                        responses.append(response)
+                except Exception:
+                    pass
+        
+        if not responses:
+            raise HTTPException(status_code=500, detail="Error: No responses generated.")
+        
+        best_response = select_best_response(responses)
+        return {
+            "best_response": best_response,
+            "all_responses": responses
+        }
+    except Exception:
+        raise HTTPException(status_code=500, detail="Internal Server Error")
 
 if __name__ == "__main__":
     uvicorn.run(app, host="0.0.0.0", port=7860)