from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel
import torch
import locale
import os

locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')

# Load model directly
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2")

# Use bitsandbytes to load the model in 8-bit precision
bnb_config = BitsAndBytesConfig(
    load_in_8bit=True,
    bnb_8bit_use_double_quant=True,
    bnb_8bit_quant_type="nf8",
    bnb_8bit_compute_dtype=torch.bfloat16
)
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2", quantization_config=bnb_config)

# Load adapter configuration
adapter_config_dir = "adapter_config"
# Load the adapters into the model
model = PeftModel.from_pretrained("mistralai/Mistral-7B-Instruct-v0.2", adapter_config=adapter_config_dir)

# FastAPI app
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel

app = FastAPI()

class Question(BaseModel):
    question: str

class Answer(BaseModel):
    answer: str

@app.post("/ask", response_model=Answer)
async def ask_question(question: Question):
    try:
        inputs = tokenizer(question.question, return_tensors="pt")
        outputs = model.generate(**inputs)
        answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
        return Answer(answer=answer)
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))

# To run the FastAPI app, use the following command:
# uvicorn app:app --host 0.0.0.0 --port 8000