Update api/utils.py

api/utils.py

@@ -1,5 +1,3 @@
-# api/utils.py
-
 from datetime import datetime
 import json
 from typing import Any, Dict, Optional
@@ -19,69 +17,13 @@ from api.models import ChatRequest
 
 from api.logger import setup_logger
 
-import uuid  # Added import for uuid
+import uuid  # Ensure this is imported if needed elsewhere
 
 logger = setup_logger(__name__)
 
-def create_chat_completion_data(
-    content: str, model: str, timestamp: int, finish_reason: Optional[str] = None
-) -> Dict[str, Any]:
-    return {
-        "id": f"chatcmpl-{uuid.uuid4()}",
-        "object": "chat.completion.chunk",
-        "created": timestamp,
-        "model": model,
-        "choices": [
-            {
-                "index": 0,
-                "delta": {"content": content, "role": "assistant"},
-                "finish_reason": finish_reason,
-            }
-        ],
-        "usage": None,
-    }
-
-def message_to_dict(message, model_prefix: Optional[str] = None):
-    if isinstance(message.content, str):
-        content = message.content
-        if model_prefix:
-            content = f"{model_prefix} {content}"
-        return {"role": message.role, "content": content}
-    elif isinstance(message.content, list) and len(message.content) == 2:
-        content = message.content[0]["text"]
-        if model_prefix:
-            content = f"{model_prefix} {content}"
-        return {
-            "role": message.role,
-            "content": content,
-            "data": {
-                "imageBase64": message.content[1]["image_url"]["url"],
-                "fileText": "",
-                "title": "snapshot",
-            },
-        }
-    else:
-        return {"role": message.role, "content": message.content}
-
-def strip_model_prefix(content: str, model_prefix: str) -> str:
-    """Remove the model prefix and language specifier from the response content if present."""
-    # Remove the model prefix
-    if content.startswith(model_prefix):
-        content = content[len(model_prefix):].strip()
-    
-    # Remove language specifier from code blocks (e.g., ```html -> ```)
-    if content.startswith("```"):
-        first_newline = content.find("\n")
-        if first_newline != -1:
-            # Retain the triple backticks and the content after the language specifier
-            content = "```\n" + content[first_newline + 1:]
-    
-    return content
-
 async def process_streaming_response(request: ChatRequest):
     agent_mode = AGENT_MODE.get(request.model, {})
     trending_agent_mode = TRENDING_AGENT_MODE.get(request.model, {})
-    model_prefix = MODEL_PREFIXES.get(request.model, "")
     referer_path = MODEL_REFERERS.get(request.model, f"/?model={request.model}")
     referer_url = f"{BASE_URL}{referer_path}"
 
@@ -90,7 +32,7 @@ async def process_streaming_response(request: ChatRequest):
     dynamic_headers['Referer'] = referer_url
 
     json_data = {
-        "messages": [message_to_dict(msg, model_prefix=model_prefix) for msg in request.messages],
+        "messages": request.messages,  # Directly pass messages without processing
         "previewToken": None,
         "userId": None,
         "codeModelMode": True,
@@ -121,18 +63,8 @@ async def process_streaming_response(request: ChatRequest):
                 timeout=100,
             ) as response:
                 response.raise_for_status()
-                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:]
-                        # Strip the model prefix and handle code blocks
-                        cleaned_content = strip_model_prefix(content, model_prefix)
-                        yield f"data: {json.dumps(create_chat_completion_data(cleaned_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"
+                async for chunk in response.aiter_text():
+                    yield chunk  # Directly yield the chunk without modification
         except httpx.HTTPStatusError as e:
             logger.error(f"HTTP error occurred: {e}")
             raise HTTPException(status_code=e.response.status_code, detail=str(e))
@@ -143,7 +75,6 @@ async def process_streaming_response(request: ChatRequest):
 async def process_non_streaming_response(request: ChatRequest):
     agent_mode = AGENT_MODE.get(request.model, {})
     trending_agent_mode = TRENDING_AGENT_MODE.get(request.model, {})
-    model_prefix = MODEL_PREFIXES.get(request.model, "")
     referer_path = MODEL_REFERERS.get(request.model, f"/?model={request.model}")
     referer_url = f"{BASE_URL}{referer_path}"
 
@@ -152,7 +83,7 @@ async def process_non_streaming_response(request: ChatRequest):
     dynamic_headers['Referer'] = referer_url
 
     json_data = {
-        "messages": [message_to_dict(msg, model_prefix=model_prefix) for msg in request.messages],
+        "messages": request.messages,  # Directly pass messages without processing
         "previewToken": None,
         "userId": None,
         "codeModelMode": True,
@@ -172,38 +103,20 @@ async def process_non_streaming_response(request: ChatRequest):
         "mobileClient": False,
         "userSelectedModel": MODEL_MAPPING.get(request.model, request.model),
     }
-    full_response = ""
+
     async with httpx.AsyncClient() as client:
         try:
-            async with client.stream(
-                method="POST", url=f"{BASE_URL}/api/chat", headers=dynamic_headers, json=json_data
-            ) as response:
-                response.raise_for_status()
-                async for chunk in response.aiter_text():
-                    full_response += chunk
+            response = await client.post(
+                f"{BASE_URL}/api/chat",
+                headers=dynamic_headers,
+                json=json_data,
+                timeout=100,
+            )
+            response.raise_for_status()
+            return response.json()  # Directly return the JSON response
         except httpx.HTTPStatusError as e:
             logger.error(f"HTTP error occurred: {e}")
             raise HTTPException(status_code=e.response.status_code, detail=str(e))
         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:]
-
-    # Strip the model prefix and handle code blocks
-    cleaned_full_response = strip_model_prefix(full_response, model_prefix)
-
-    return {
-        "id": f"chatcmpl-{uuid.uuid4()}",
-        "object": "chat.completion",
-        "created": int(datetime.now().timestamp()),
-        "model": request.model,
-        "choices": [
-            {
-                "index": 0,
-                "message": {"role": "assistant", "content": cleaned_full_response},
-                "finish_reason": "stop",
-            }
-        ],
-        "usage": None,
-    }