Update api/utils.py

api/utils.py

@@ -3,7 +3,6 @@ import json
 import uuid
 import asyncio
 import random
-import string
 from typing import Any, Dict, Optional
 
 import httpx
@@ -60,14 +59,6 @@ def message_to_dict(message, model_prefix: Optional[str] = None):
         }
     return {"role": message.role, "content": content}
 
-# Function to strip model prefix from content if present
-def strip_model_prefix(content: str, model_prefix: Optional[str] = None) -> str:
-    """Remove the model prefix from the response content if present."""
-    if model_prefix and content.startswith(model_prefix):
-        logger.debug(f"Stripping prefix '{model_prefix}' from content.")
-        return content[len(model_prefix):].strip()
-    return content
-
 # Process streaming response with headers from config.py
 async def process_streaming_response(request: ChatRequest):
     logger.info(f"Processing streaming response for Model: {request.model}")
@@ -121,11 +112,8 @@ async def process_streaming_response(request: ChatRequest):
                 async for line in response.aiter_lines():
                     timestamp = int(datetime.now().timestamp())
                     if line:
-                        content = line
-                        if content.startswith("$@$v=undefined-rv1$@$"):
-                            content = content[21:]
-                        cleaned_content = strip_model_prefix(content, model_prefix)
-                        yield f"data: {json.dumps(create_chat_completion_data(cleaned_content, request.model, timestamp))}\n\n"
+                        # Directly yield each line without additional processing
+                        yield f"data: {json.dumps(create_chat_completion_data(line, request.model, timestamp))}\n\n"
 
                 yield f"data: {json.dumps(create_chat_completion_data('', request.model, timestamp, 'stop'))}\n\n"
                 yield "data: [DONE]\n\n"
@@ -191,10 +179,6 @@ async def process_non_streaming_response(request: ChatRequest):
         except httpx.RequestError as e:
             logger.error(f"Error occurred during request: {e}")
             raise HTTPException(status_code=500, detail=str(e))
-    if full_response.startswith("$@$v=undefined-rv1$@$"):
-        full_response = full_response[21:]
-
-    cleaned_full_response = strip_model_prefix(full_response, model_prefix)
 
     return {
         "id": f"chatcmpl-{uuid.uuid4()}",
@@ -204,7 +188,7 @@ async def process_non_streaming_response(request: ChatRequest):
         "choices": [
             {
                 "index": 0,
-                "message": {"role": "assistant", "content": cleaned_full_response},
+                "message": {"role": "assistant", "content": full_response},
                 "finish_reason": "stop",
             }
         ],