Update api/utils.py

api/utils.py

@@ -24,7 +24,6 @@ from api.logger import (
     log_model_delay,
     log_http_error,
     log_request_error,
-    log_strip_prefix,
 )
 
 # Helper function to generate a random alphanumeric chat ID
@@ -71,15 +70,7 @@ def message_to_dict(message, model_prefix: Optional[str] = None):
     else:
         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):
-        log_strip_prefix(model_prefix, content)
-        return content[len(model_prefix):].strip()
-    return content
-
-# Streaming response processing with headers from config.py
+# Streaming response processing
 async def process_streaming_response(request: ChatRequest):
     chat_id = generate_chat_id() if request.model in MODEL_REFERERS else None
     referer_path = MODEL_REFERERS.get(request.model, "")
@@ -92,10 +83,8 @@ async def process_streaming_response(request: ChatRequest):
     trending_agent_mode = TRENDING_AGENT_MODE.get(request.model, {})
     model_prefix = MODEL_PREFIXES.get(request.model, "")
 
-    # Generate headers for API chat request with dynamic Referer
     headers_api_chat = get_headers_api_chat(referer_url)
 
-    # Introduce delay for 'o1-preview' model
     if request.model == 'o1-preview':
         delay_seconds = random.randint(20, 60)
         log_model_delay(delay_seconds, request.model, chat_id)
@@ -137,10 +126,8 @@ async def process_streaming_response(request: ChatRequest):
                 response.raise_for_status()
                 async for line in response.aiter_lines():
                     timestamp = int(datetime.now().timestamp())
-                    if line:
-                        content = line.lstrip("$@$v=undefined-rv1$@$")
-                        cleaned_content = strip_model_prefix(content, model_prefix)
-                        yield f"data: {json.dumps(create_chat_completion_data(cleaned_content, request.model, timestamp))}\n\n"
+                    content = line.lstrip("$@$v=undefined-rv1$@$")  # Trim only if needed
+                    yield f"data: {json.dumps(create_chat_completion_data(content, 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"
@@ -151,24 +138,21 @@ async def process_streaming_response(request: ChatRequest):
             log_request_error(e, chat_id)
             raise HTTPException(status_code=500, detail=str(e))
 
-# Non-streaming response processing with headers from config.py
+# Non-streaming response processing
 async def process_non_streaming_response(request: ChatRequest):
     chat_id = generate_chat_id() if request.model in MODEL_REFERERS else None
     referer_path = MODEL_REFERERS.get(request.model, "")
     referer_url = f"{BASE_URL}/chat/{chat_id}?model={request.model}" if chat_id else BASE_URL
 
-    # Log with chat ID, model, and referer URL if applicable
     log_generated_chat_id_with_referer(chat_id, request.model, referer_url)
 
     agent_mode = AGENT_MODE.get(request.model, {})
     trending_agent_mode = TRENDING_AGENT_MODE.get(request.model, {})
     model_prefix = MODEL_PREFIXES.get(request.model, "")
 
-    # Generate headers for API chat request and secondary chat request with dynamic Referer
     headers_api_chat = get_headers_api_chat(referer_url)
     headers_chat = get_headers_chat(referer_url, next_action=str(uuid.uuid4()), next_router_state_tree=json.dumps([""]))
 
-    # Introduce delay for 'o1-preview' model
     if request.model == 'o1-preview':
         delay_seconds = random.randint(20, 60)
         log_model_delay(delay_seconds, request.model, chat_id)
@@ -214,11 +198,7 @@ async def process_non_streaming_response(request: ChatRequest):
             log_request_error(e, chat_id)
             raise HTTPException(status_code=500, detail=str(e))
     
-    if full_response.startswith("$@$v=undefined-rv1$@$"):
-        full_response = full_response[21:]
-
-    # Strip the model prefix from the full response
-    cleaned_full_response = strip_model_prefix(full_response, model_prefix)
+    full_response = full_response.lstrip("$@$v=undefined-rv1$@$")  # Trim only if needed
 
     return {
         "id": f"chatcmpl-{uuid.uuid4()}",
@@ -228,7 +208,7 @@ 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",
             }
         ],