import os
import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

from huggingface_hub import login
login(token=os.environ["HF_TOKEN"])

model_id = "mistralai/Mistral-7B-Instruct-v0.2"

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    torch_dtype=torch.float16,
)

# Prompt formatter
def build_prompt(user_input):
    return f"<s>[INST] {user_input.strip()} [/INST]"

# Chat function
def chat(user_input):
    prompt = build_prompt(user_input)
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

    outputs = model.generate(
        **inputs,
        max_new_tokens=300,
        do_sample=True,
        temperature=0.7,
        top_p=0.95,
        top_k=50
    )

    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return response.split('[/INST]')[-1].strip()

# Gradio Interface
demo = gr.Interface(fn=chat, inputs="text", outputs="text", title="Mistral AI Assistant")
demo.launch()