Create app.py

app.py

@@ -0,0 +1,91 @@
+import os
+from fastapi import FastAPI, HTTPException
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import BaseModel
+from typing import List
+import httpx
+import logging
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+# Get HF token from Space Secrets
+HF_TOKEN = os.environ.get("HF_TOKEN")
+if not HF_TOKEN:
+    raise RuntimeError("HF_TOKEN environment variable not set")
+
+app = FastAPI(title="HF Inference Proxy")
+
+# Enable CORS for Unity
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+class Message(BaseModel):
+    role: str
+    content: str
+
+class InferenceRequest(BaseModel):
+    model: str
+    temperature: float
+    messages: List[Message]
+    max_tokens: int = 512
+
+@app.get("/")
+async def root():
+    return {
+        "status": "running",
+        "message": "HF Inference Proxy Active",
+        "endpoints": {
+            "extract": "POST /extract"
+        }
+    }
+
+@app.post("/extract")
+async def proxy_inference(request: InferenceRequest):
+    try:
+        logger.info(f"Request: {request.model}, {len(request.messages)} messages")
+        
+        async with httpx.AsyncClient(timeout=120.0) as client:
+            hf_response = await client.post(
+                "https://router.huggingface.co/v1/chat/completions",
+                json={
+                    "model": f"{request.model}:fastest",
+                    "temperature": request.temperature,
+                    "max_tokens": request.max_tokens,
+                    "messages": [
+                        {"role": msg.role, "content": msg.content} 
+                        for msg in request.messages
+                    ]
+                },
+                headers={
+                    "Authorization": f"Bearer {HF_TOKEN}",
+                    "Content-Type": "application/json"
+                }
+            )
+            
+            if hf_response.status_code != 200:
+                logger.error(f"HF API error: {hf_response.text}")
+                raise HTTPException(
+                    status_code=hf_response.status_code,
+                    detail=hf_response.text
+                )
+            
+            result = hf_response.json()
+            logger.info("Request successful")
+            return result
+            
+    except httpx.HTTPError as e:
+        logger.error(f"HTTP error: {str(e)}")
+        raise HTTPException(status_code=500, detail=f"Proxy error: {str(e)}")
+    except Exception as e:
+        logger.error(f"Error: {str(e)}")
+        raise HTTPException(status_code=500, detail=f"Server error: {str(e)}")
+
+@app.get("/health")
+async def health():
+    return {"status": "healthy", "hf_configured": bool(HF_TOKEN)}