Update app.py

app.py

@@ -1,22 +1,24 @@
 from fastapi import FastAPI, Request
 from fastapi.responses import JSONResponse, FileResponse
 from transformers import pipeline
-from huggingface_hub import login
 import os
 
-# 🛠️ Fix write permission issue
-os.environ["HF_HOME"] = "/data"  # ✅ Hugging Face model cache path
+# ✅ Use /data for HF cache — only writable directory in Spaces
+os.environ["HF_HOME"] = "/data"
 
-# 🔐 Login with token from Hugging Face Secrets (don't hardcode!)
+# ✅ Get token and model from env
 hf_token = os.getenv("HF_TOKEN")
-if hf_token:
-    login(token=hf_token)
-
-# ✅ Load the model pipeline (CPU-friendly)
 model_id = os.getenv("MODEL_ID", "mistralai/Mistral-7B-Instruct-v0.2")
-pipe = pipeline("text-generation", model=model_id, trust_remote_code=True)
 
-# 🚀 FastAPI app
+# ✅ Load pipeline without login(), using token directly
+pipe = pipeline(
+    "text-generation",
+    model=model_id,
+    token=hf_token,
+    trust_remote_code=True
+)
+
+# ✅ Set up FastAPI app
 app = FastAPI()
 
 @app.get("/")
@@ -28,13 +30,13 @@ async def serve_script():
     return FileResponse("script.js")
 
 @app.post("/api")
-async def chat_endpoint(request: Request):
+async def ask_ai(request: Request):
     data = await request.json()
     question = data.get("question", "")
     prompt = f"[INST] {question.strip()} [/INST]"
     try:
-        result = pipe(prompt, max_new_tokens=256, temperature=0.7)
-        response = result[0]["generated_text"].split("[/INST]")[-1].strip()
+        output = pipe(prompt, max_new_tokens=256, temperature=0.7)
+        response = output[0]["generated_text"].split("[/INST]")[-1].strip()
         return JSONResponse({"answer": response})
     except Exception as e:
         return JSONResponse({"answer": f"Error: {str(e)}"})