inshallah3

routers/mistral.py

@@ -6,6 +6,7 @@ from mistralai.client import Mistral
 from auth import verify_token
 import os
 from schemas.common import APIResponse, ChatMessage, ChatMemoryRequest # Import common schemas
+from routers.gemini import generate_gemini_chat_response, GEMINI_DEFAULT_CHAT_MODEL, ALLOWED_CHAT_MODELS
 import asyncio
 import time
 from functools import wraps
@@ -134,32 +135,32 @@ async def mistral_chat(request: LLMRequest, token: Annotated[str, Depends(verify
         print(f"Error calling Mistral chat completion: {e}")
         raise HTTPException(status_code=500, detail=f"Mistral API error: {str(e)}")
 
-@router.post("/chat-with-memory", response_model=APIResponse)
-async def mistral_chat_with_memory(request: ChatMemoryRequest, token: Annotated[str, Depends(verify_token)]):    
+@router.post("/chat-with-memory", response_model=APIResponse, deprecated=True)
+async def mistral_chat_with_memory(request: ChatMemoryRequest, token: Annotated[str, Depends(verify_token)]):
+    """Backward-compatible alias rerouted to Gemini (Mistral is rate-limited).
+
+    Keeps the same path/schema so existing clients don't change.
+    New clients should use POST /gemini/chat-with-memory directly.
+    Mistral model names are mapped to GEMINI_CHAT_MODEL (default gemini-flash-latest);
+    a request model matching one of ALLOWED_CHAT_MODELS is passed through.
+    """
     try:
-        async def chat_with_memory_call(client, request):
-            # Convert Pydantic models to dicts for the Mistral API call
-            messages_dict = [msg.model_dump() for msg in request.messages]
-            
-            return await client.chat.complete_async(
-                model=request.model,
-                messages=messages_dict,
-            )
-        
-        response = await mistral_manager.rate_limited_call(chat_with_memory_call, request)
-        
-        if response.choices and response.choices[0].message:
-            content = response.choices[0].message.content
-            return APIResponse(success=True, data={"response": content})
+        requested = (request.model or "").strip()
+        if requested in ALLOWED_CHAT_MODELS:
+            model = requested
         else:
-            return APIResponse(success=False, error="No response content received from Mistral.", data=response.dict())
-            
+            if requested:
+                print(f"Mistral model '{requested}' rerouted to Gemini model '{GEMINI_DEFAULT_CHAT_MODEL}'")
+            model = GEMINI_DEFAULT_CHAT_MODEL
+        text = await asyncio.to_thread(generate_gemini_chat_response, model, request.messages)
+        return APIResponse(success=True, data={"response": text})
+
     except HTTPException:
         # Re-raise HTTP exceptions (like timeout)
         raise
     except Exception as e:
-        print(f"Error calling Mistral chat completion with memory: {e}")
-        raise HTTPException(status_code=500, detail=f"Mistral API error: {str(e)}")
+        print(f"Error calling rerouted (Gemini) chat completion with memory: {e}")
+        raise HTTPException(status_code=500, detail=f"Chat API error: {str(e)}")
 
 @router.get("/health", response_model=APIResponse)
 async def mistral_health(token: Annotated[str, Depends(verify_token)]):