app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

MODEL_NAME = "Abeersherif/Medical_Homework2"

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
)

def respond(message, history, system_message, max_tokens, temperature, top_p):
    """
    message: str                    -> latest user message
    history: list[[user, bot], ...] -> previous chat turns (default tuples mode)
    """

    # Build a simple text prompt (no chat template)
    conversation = f"System: {system_message}\n\n"
    for user_msg, bot_msg in history:
        conversation += f"User: {user_msg}\nAssistant: {bot_msg}\n"
    conversation += f"User: {message}\nAssistant:"

    result = pipe(
        conversation,
        max_new_tokens=int(max_tokens),
        temperature=float(temperature),
        top_p=float(top_p),
        do_sample=True,
    )[0]["generated_text"]

    # Keep only what the assistant said last
    if "Assistant:" in result:
        result = result.split("Assistant:")[-1]

    return result.strip()


chatbot = gr.ChatInterface(
    fn=respond,
    # ⚠️ use default history format (tuples), do NOT set type="messages" here
    additional_inputs=[
        gr.Textbox(
            "You are a helpful medical assistant. Answer concisely with brief reasoning.",
            label="System message",
        ),
        gr.Slider(1, 512, value=256, step=1, label="Max new tokens"),
        gr.Slider(0.1, 4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p"),
    ],
)

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