app.py

import os
import logging
import tempfile
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel

TMP_CACHE = os.environ.get("HF_CACHE_DIR", os.path.join(tempfile.gettempdir(), "hf_cache"))
try:
    os.makedirs(TMP_CACHE, exist_ok=True)
except Exception:
    TMP_CACHE = tempfile.gettempdir()

os.environ["TRANSFORMERS_CACHE"] = TMP_CACHE
os.environ["HF_HOME"] = TMP_CACHE
os.environ["HF_DATASETS_CACHE"] = TMP_CACHE
os.environ["HF_METRICS_CACHE"] = TMP_CACHE

app = FastAPI(title="DirectEd LoRA API")

class PromptRequest(BaseModel):
    prompt: str
    max_new_tokens: int = 2048

@app.get("/")
def root():
    return {"status": "AI backend is running"}

pipe = None

@app.on_event("startup")
def load_model():
    global pipe
    try:
        from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
        from peft import PeftModel

        BASE_MODEL = "unsloth/llama-3-8b-Instruct-bnb-4bit"
        ADAPTER_REPO = "rayymaxx/DirectEd-AI-LoRA"

        tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
        base_model = AutoModelForCausalLM.from_pretrained(
            BASE_MODEL,
            device_map="auto",
            torch_dtype="auto",
        )

        model = PeftModel.from_pretrained(base_model, ADAPTER_REPO)
        model.eval()

        pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, device_map="auto")
        logging.info("Model and adapter loaded successfully.")

    except Exception as e:
        logging.exception("Failed to load model at startup: %s", e)
        pipe = None

@app.post("/generate")
def generate(req: PromptRequest):
    if pipe is None:
        raise HTTPException(status_code=503, detail="Model not loaded. Check logs.")

    try:
        output = pipe(req.prompt, max_new_tokens=req.max_new_tokens, do_sample=True, temperature=0.7)
        text = output[0].get("generated_text", "").strip()

        if text.startswith(req.prompt):
            text = text[len(req.prompt):].strip()

        if not text:
            text = "No response generated by the model."

        return {"response": text}

    except Exception as e:
        logging.exception("Generation failed for prompt '%s': %s", req.prompt, e)
        raise HTTPException(status_code=500, detail=f"Generation failed: {e}")