main.py

from fastapi import FastAPI, Request
from fastapi.responses import StreamingResponse
from huggingface_hub import InferenceClient
import json
import os

app = FastAPI()

# تأكد من وضع HF_TOKEN في الـ Secrets في إعدادات الـ Space
client = InferenceClient(api_key=os.getenv("HF_TOKEN"))

@app.get("/")
def read_root():
    return {"status": "DEV_STOREOMALL_V10_ACTIVE"}

@app.post("/v1/chat/completions")
async def chat_completions(request: Request):
    data = await request.json()
    
    def stream_generator():
        # استدعاء موديل Qwen 3 Coder 40B أو Qwen 2.5
        stream = client.chat_completion(
            model="Qwen/Qwen2.5-Coder-32B-Instruct", 
            messages=data["messages"],
            max_tokens=data.get("max_tokens", 3000),
            temperature=data.get("temperature", 0.7),
            stream=True
        )
        
        for chunk in stream:
            # تنسيق البيانات ليتوافق مع كود الـ PHP الخاص بك
            yield f"data: {json.dumps(chunk)}\n\n"
        yield "data: [DONE]\n\n"

    return StreamingResponse(stream_generator(), media_type="text/event-stream")