# app.py
# =============
# This is a complete app.py file for a text generation app using the Qwen/Qwen2.5-Coder-0.5B-Instruct model.
# The app uses the Gradio library to create a web interface for interacting with the model.

# Imports
# =======
import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

# Constants
# =========
MODEL_NAME = "Qwen/Qwen2.5-Coder-0.5B-Instruct"
SYSTEM_MESSAGE = "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."

# Load Model and Tokenizer
# ========================
def load_model_and_tokenizer():
    """
    Load the model and tokenizer from Hugging Face.
    """
    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
    model = AutoModelForCausalLM.from_pretrained(
        MODEL_NAME,
        torch_dtype="auto",
        device_map="cpu"  # Ensure the model runs on the CPU
    )
    return model, tokenizer

model, tokenizer = load_model_and_tokenizer()

# Generate Response
# =================
def generate_response(prompt, chat_history):
    """
    Generate a response from the model based on the user prompt and chat history.
    """
    messages = [{"role": "system", "content": SYSTEM_MESSAGE}] + chat_history + [{"role": "user", "content": prompt}]
    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )
    model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

    generated_ids = model.generate(
        **model_inputs,
        max_new_tokens=512,
        do_sample=True,
        top_k=50,
        top_p=0.95,
        temperature=0.7,
        stream=True
    )

    response = ""
    for new_token in generated_ids[0][len(model_inputs.input_ids[0]):]:
        response += tokenizer.decode([new_token], skip_special_tokens=True)
        yield response

# Clear Chat History
# ==================
def clear_chat():
    """
    Clear the chat history.
    """
    return [], []

# Gradio Interface
# =================
def gradio_interface():
    """
    Create and launch the Gradio interface.
    """
    with gr.Blocks() as demo:
        chatbot = gr.Chatbot(label="Chat with Qwen/Qwen2.5-Coder-0.5B-Instruct")
        msg = gr.Textbox(label="User Input")
        clear = gr.Button("Clear Chat")

        def respond(message, chat_history):
            chat_history.append({"role": "user", "content": message})
            response = generate_response(message, chat_history)
            chat_history.append({"role": "assistant", "content": response})
            return chat_history, chat_history

        msg.submit(respond, [msg, chatbot], [chatbot, chatbot])
        clear.click(clear_chat, None, [chatbot, chatbot])

    demo.launch()

# Main
# ====
if __name__ == "__main__":
    gradio_interface()

# Dependencies
# =============
# The following dependencies are required to run this app:
# - transformers
# - gradio
# - torch
#
# You can install these dependencies using pip:
# pip install transformers gradio torch