app.py

# app.py
import torch
import gradio as gr
import spaces
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
from peft import PeftModel

# ────────────────────────────────────────────────────────────────
BASE_MODEL = "unsloth/Phi-3-mini-4k-instruct-bnb-4bit"
LORA_PATH = "saadkhi/SQL_Chat_finetuned_model"

MAX_NEW_TOKENS = 180
TEMPERATURE = 0.0
DO_SAMPLE = False

print("Loading quantized base model on CPU...")
print("(GPU will be used only during inference if available)")

# 4-bit quantization config
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16,
    bnb_4bit_use_double_quant=True,
)

# Load base model → always on CPU first
model = AutoModelForCausalLM.from_pretrained(
    BASE_MODEL,
    quantization_config=bnb_config,
    device_map="cpu",
    trust_remote_code=True,
    torch_dtype=torch.bfloat16,
)

print("Loading LoRA adapters...")
model = PeftModel.from_pretrained(model, LORA_PATH)

# Merge for faster inference (very recommended)
print("Merging LoRA into base model...")
model = model.merge_and_unload()

# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
tokenizer.pad_token = tokenizer.eos_token

model.eval()
# ────────────────────────────────────────────────────────────────

@spaces.GPU(duration=60, max_requests=20)  # safe values for ZeroGPU
def generate_sql(prompt: str):
    # Prepare chat format
    messages = [
        {"role": "user", "content": prompt}
    ]

    # Tokenize on CPU (safe everywhere)
    inputs = tokenizer.apply_chat_template(
        messages,
        tokenize=True,
        add_generation_prompt=True,
        return_tensors="pt"
    )

    # Choose device dynamically - this is the ZeroGPU-safe way
    device = "cuda" if torch.cuda.is_available() else "cpu"
    print(f"→ Running inference on device: {device}")

    inputs = inputs.to(device)

    with torch.inference_mode():
        outputs = model.generate(
            input_ids=inputs,
            max_new_tokens=MAX_NEW_TOKENS,
            temperature=TEMPERATURE,
            do_sample=DO_SAMPLE,
            pad_token_id=tokenizer.eos_token_id,
            eos_token_id=tokenizer.eos_token_id,
        )

    # Decode and clean output
    response = tokenizer.decode(outputs[0], skip_special_tokens=True)

    # Remove user's prompt + assistant tag if present
    if "<|assistant|>" in response:
        response = response.split("<|assistant|>", 1)[-1].strip()

    # Cut at end token if exists
    if "<|end|>" in response:
        response = response.split("<|end|>", 1)[0].strip()

    return response.strip()


# ────────────────────────────────────────────────────────────────
demo = gr.Interface(
    fn=generate_sql,
    inputs=gr.Textbox(
        label="Ask a question about SQL",
        placeholder="Delete duplicate rows from users table based on email",
        lines=3,
    ),
    outputs=gr.Textbox(label="Generated SQL Query"),
    title="SQL Chatbot – Phi-3-mini + LoRA",
    description=(
        "Fine-tuned Phi-3-mini-4k-instruct (4bit) for generating SQL queries\n\n"
        "Works on ZeroGPU and regular GPU hardware"
    ),
    examples=[
        ["Find duplicate emails in users table"],
        ["Top 5 highest paid employees"],
        ["Count orders per customer last month"],
        ["Show all products that haven't been ordered in the last 6 months"],
        ["Update all orders from 2024 to status 'completed'"],
    ],
    cache_examples=False,
)

if __name__ == "__main__":
    demo.launch()