import gradio as gr
from huggingface_hub import InferenceClient

client = InferenceClient("Qwen/Qwen2.5-7B-Instruct")

def respond(message, history):
    
    messages = [{"role": "system", "content": "You're really good at recommending books based on what genres and tropes the user likes. Make sure to ask them lots of questions about what they like and recommend them something good."}]
    
    if history:
        messages.extend(history)
        
    messages.append({"role": "user", "content": message})
    
    response_text = ""
    response = client.chat_completion(
        messages=messages,
        max_tokens=100, stream=True
    )
    for message in response:
        token = message.choices[0].delta.content
        response_text += token
        yield response_text

chatbot = gr.ChatInterface(respond)

chatbot.launch()