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

# ────────────────────────────────────────────────────────────────
#   Configuration - fastest practical settings
# ────────────────────────────────────────────────────────────────

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

MAX_NEW_TOKENS = 180          # ← keep reasonable
TEMPERATURE    = 0.0          # greedy = fastest & most deterministic
DO_SAMPLE      = False        # no sampling = faster

# ────────────────────────────────────────────────────────────────
#   4-bit quantization config (this is the key speedup)
# ────────────────────────────────────────────────────────────────

bnb_config = BitsAndBytesConfig(
    load_in_4bit              = True,
    bnb_4bit_quant_type       = "nf4",           # "nf4" usually fastest + good quality
    bnb_4bit_use_double_quant = True,            # nested quantization → extra memory saving
    bnb_4bit_compute_dtype    = torch.bfloat16   # fastest compute type on modern GPUs
)

print("Loading quantized base model...")
model = AutoModelForCausalLM.from_pretrained(
    BASE_MODEL,
    quantization_config = bnb_config,
    device_map          = "auto",                # auto = best available (cuda > cpu)
    trust_remote_code   = True,
    torch_dtype         = torch.bfloat16
)

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

# Important: merge LoRA weights into base (faster inference, less overhead)
model = model.merge_and_unload()

tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)

# Optional: small speedup boost on supported hardware
if torch.cuda.is_available():
    try:
        import torch.backends.cuda
        torch.backends.cuda.enable_flash_sdp(True)  # flash scaled dot product
    except:
        pass

model.eval()
print("Model ready!")

# ────────────────────────────────────────────────────────────────
def generate_sql(prompt: str):
    # Use proper chat template (Phi-3 expects it)
    messages = [{"role": "user", "content": prompt}]
    
    inputs = tokenizer.apply_chat_template(
        messages,
        tokenize           = True,
        add_generation_prompt = True,
        return_tensors     = "pt"
    ).to(model.device)

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

    response = tokenizer.decode(outputs[0], skip_special_tokens=True)

    # Clean output - try to get only assistant's answer
    if "<|assistant|>" in response:
        response = response.split("<|assistant|>", 1)[-1].strip()
    response = response.split("<|end|>")[0].strip()

    return response

# ────────────────────────────────────────────────────────────────
demo = gr.Interface(
    fn           = generate_sql,
    inputs       = gr.Textbox(
        label       = "Ask SQL related question",
        placeholder = "Show me all employees with salary > 50000...",
        lines       = 3
    ),
    outputs      = gr.Textbox(label="Generated SQL / Answer"),
    title        = "SQL Chatbot - Fast Version",
    description  = "Phi-3-mini 4bit quantized + LoRA",
    examples     = [
        ["Find duplicate emails in users table"],
        ["Top 5 highest paid employees"],
        ["Count orders per customer last month"]
    ],
    allow_flagging = "never"
)

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