app.py

# app.py

import os
import logging
from http.client import HTTPMessage
from http import HTTPStatus
import gradio as gr
from dashscope import Generation, Role
from dashscope.api_entities.dashscope_response import DashScopeResponse
from typing import List, Optional, Tuple, Dict
from urllib.error import HTTPError

# Configure logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

# Environment variable for API token
YOUR_API_TOKEN = os.getenv('YOUR_API_TOKEN')
if not YOUR_API_TOKEN:
    logging.error("API token not found in environment variables.")
    exit(1)

dashscope.api_key = YOUR_API_TOKEN

History = List[Tuple[str, str]]
Messages = List[Dict[str, str]]

default_system = 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.'

def clear_session() -> History:
    return '', []

def modify_system_session(system: str) -> Tuple[str, str, History]:
    if not system:
        system = default_system
    return system, system, []

def history_to_messages(history: History, system: str) -> Messages:
    messages = [{'role': Role.SYSTEM, 'content': system}]
    for user_msg, assistant_msg in history:
        messages.append({'role': Role.USER, 'content': user_msg})
        messages.append({'role': Role.ASSISTANT, 'content': assistant_msg})
    return messages

def messages_to_history(messages: Messages) -> Tuple[str, History]:
    assert messages[0]['role'] == Role.SYSTEM
    system = messages[0]['content']
    history = [(msg['content'], next_msg['content']) for msg, next_msg in zip(messages[1::2], messages[2::2])]
    return system, history

def model_chat(query: Optional[str], history: Optional[History], system: str, radio: str) -> Tuple[str, str, History]:
    if not query:
        query = ''
    if not history:
        history = []
    messages = history_to_messages(history, system)
    messages.append({'role': Role.USER, 'content': query})
    label_model = f"qwen2.5-coder-{radio.lower()}-instruct"
    try:
        gen = Generation.call(
            model=label_model,
            messages=messages,
            result_format='message',
            stream=True
        )
        for response in gen:
            if response.status_code == HTTPStatus.OK:
                role = response.output.choices[0].message.role
                content = response.output.choices[0].message.content
                system, history = messages_to_history(messages + [{'role': role, 'content': content}])
                yield '', history, system
            else:
                raise HTTPError(code=response.status_code, msg=f'Request ID: {response.request_id}, Status Code: {response.status_code}, Error Code: {response.code}, Error Message: {response.message}', hdrs=HTTPMessage(), url='http://example.com', fp=None)
    except HTTPError as e:
        logging.error(f"HTTP Error: {e}")
        yield '', history, system
    except Exception as e:
        logging.error(f"Unexpected error: {e}")
        yield '', history, system

def choose_radio(radio, system):
    mark_ = gr.Markdown(value=f"<center><font size=8>Qwen2.5-Coder-{radio}-instruct👾</center>")
    chatbot = gr.Chatbot(label=f'Qwen2.5-Coder-{radio.lower()}-instruct')
    if not system:
        system = default_system
    return mark_, chatbot, system, system, ""

def update_other_radios(value, other_radio1, other_radio2):
    if not value:
        if other_radio1:
            selected = other_radio1
        else:
            selected = other_radio2
        return selected, other_radio1, other_radio2
    return value, "", ""

def main():
    # Create two tabs
    with gr.Blocks() as demo:
        with gr.Row():
            options_coder = ["0.5B", "1.5B", "3B", "7B", "14B", "32B"]
            with gr.Row():
                radio = gr.Radio(choices=options_coder, label="Qwen2.5-Coder:", value="32B")

        with gr.Row():
            with gr.Accordion():
                mark_ = gr.Markdown("""<center><font size=8>Qwen2.5-Coder-32B-Instruct Bot👾</center>""")
                with gr.Row():
                    with gr.Column(scale=3):
                        system_input = gr.Textbox(value=default_system, lines=1, label='System')
                    with gr.Column(scale=1):
                        modify_system = gr.Button("🛠️ Set system prompt and clear history", scale=2)
                    system_state = gr.Textbox(value=default_system, visible=False)
                chatbot = gr.Chatbot(label='Qwen2.5-Coder-32B-Instruct')
                textbox = gr.Textbox(lines=1, label='Input')
                
                with gr.Row():
                    clear_history = gr.Button("🧹 Clear History")
                    submit = gr.Button("🚀 Send")
                
                textbox.submit(model_chat,
                               inputs=[textbox, chatbot, system_state, radio],
                               outputs=[textbox, chatbot, system_input])
                submit.click(model_chat,
                             inputs=[textbox, chatbot, system_state, radio],
                             outputs=[textbox, chatbot, system_input],
                             concurrency_limit=100)
                clear_history.click(fn=clear_session,
                                    inputs=[],
                                    outputs=[textbox, chatbot])
                modify_system.click(fn=modify_system_session,
                                    inputs=[system_input],
                                    outputs=[system_state, system_input, chatbot])
        
        radio.change(choose_radio,
                     inputs=[radio, system_input],
                     outputs=[mark_, chatbot, system_state, system_input, textbox])
    
    demo.queue(api_open=False, default_concurrency_limit=40)
    demo.launch(max_threads=5, server_name="0.0.0.0", server_port=7860, share=True)

if __name__ == "__main__":
    main()