Update app.py

app.py

@@ -1,52 +1,73 @@
-from fastapi import FastAPI, Request
+from fastapi import FastAPI
 from pydantic import BaseModel
+from typing import List, Literal, Optional
 from huggingface_hub import InferenceClient
-import uvicorn
-import time
+from fastapi.responses import JSONResponse
 import uuid
-
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
+import time
+import uvicorn
 
 app = FastAPI()
+client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+# OpenAI-compatible request message
 class Message(BaseModel):
-    role: str
+    role: Literal["system", "user", "assistant"]
     content: str
 
-class ChatRequest(BaseModel):
+# OpenAI-compatible request body
+class ChatCompletionRequest(BaseModel):
+    model: str = "zephyr-7b-beta"
+    messages: List[Message]
+    temperature: Optional[float] = 0.7
+    top_p: Optional[float] = 0.95
+    max_tokens: Optional[int] = 512
+    stream: Optional[bool] = False
+
+# OpenAI-compatible response message
+class Choice(BaseModel):
+    index: int
+    message: Message
+    finish_reason: Optional[str] = "stop"
+
+# OpenAI-compatible full response
+class ChatCompletionResponse(BaseModel):
+    id: str
+    object: str = "chat.completion"
+    created: int
     model: str
-    messages: list[Message]
-    temperature: float = 0.7
-    top_p: float = 0.95
-    max_tokens: int = 512
-    stream: bool = False
+    choices: List[Choice]
 
-@app.post("/v1/chat/completions")
-async def chat(request: ChatRequest):
-    # Convert to HF-style messages
-    formatted_messages = request.messages
+@app.post("/v1/chat/completions", response_model=ChatCompletionResponse)
+async def chat_completions(request: ChatCompletionRequest):
+    # Build HuggingFace-style message list
+    messages = [{"role": m.role, "content": m.content} for m in request.messages]
 
+    # Generate chat completion
     response_text = ""
     for chunk in client.chat_completion(
-        messages=[m.model_dump() for m in formatted_messages],
+        messages,
         max_tokens=request.max_tokens,
         temperature=request.temperature,
         top_p=request.top_p,
-        stream=False  # Hugging Face Spaces don't support streaming well
+        stream=False,
     ):
-        response_text += chunk.choices[0].delta.content or ""
-
-    return {
-        "id": f"chatcmpl-{uuid.uuid4()}",
-        "object": "chat.completion",
-        "created": int(time.time()),
-        "model": request.model,
-        "choices": [
-            {
-                "index": 0,
-                "message": {"role": "assistant", "content": response_text},
-                "finish_reason": "stop",
-            }
+        response_text += chunk.choices[0].delta.content
+
+    # Build OpenAI-style response
+    chat_response = ChatCompletionResponse(
+        id=f"chatcmpl-{uuid.uuid4().hex}",
+        created=int(time.time()),
+        model=request.model,
+        choices=[
+            Choice(
+                index=0,
+                message=Message(role="assistant", content=response_text),
+            )
         ]
-    }
+    )
+    return JSONResponse(content=chat_response.dict())
 
+# Run this file directly
+if __name__ == "__main__":
+    uvicorn.run("app:app", host="52.44.98.165", port=8000, reload=True)