app.py

import os
import json
import gradio as gr
import requests
import csv
import argparse
import shutil
from vlog4chat import Vlogger4chat
from vlog4debate import Debate
from utils import download_video

#prompt_templates = {"Default ChatGPT": ""}

parser = argparse.ArgumentParser()
parser.add_argument('--video_path', default='./BV11H4y1F7uH.mp4')
parser.add_argument('--alpha', default=10, type=int, help='Determine the maximum segment number for KTS algorithm, the larger the value, the fewer segments.')
parser.add_argument('--beta', default=1, type=int, help='The smallest time gap between successive clips, in seconds.')
parser.add_argument('--data_dir', default='./', type=str, help='Directory for saving videos and logs.')
parser.add_argument('--tmp_dir', default='./', type=str, help='Directory for saving intermediate files.')

# * Models settings *
parser.add_argument('--openai_api_key', default='xxx', type=str, help='OpenAI API key')
parser.add_argument('--image_caption', action='store_true', dest='image_caption', default=True, help='Set this flag to True if you want to use BLIP Image Caption')
parser.add_argument('--dense_caption', action='store_true', dest='dense_caption', default=True, help='Set this flag to True if you want to use Dense Caption')
parser.add_argument('--feature_extractor', default='./clip-vit-base-patch32', help='Select the feature extractor model for video segmentation')
parser.add_argument('--feature_extractor_device', choices=['cuda', 'cpu'], default='cuda', help='Select the device: cuda or cpu')
parser.add_argument('--image_captioner', choices=['blip2-opt', 'blip2-flan-t5', 'blip'], dest='captioner_base_model', default='blip2-opt', help='blip2 requires 15G GPU memory, blip requires 6G GPU memory')
parser.add_argument('--image_captioner_device', choices=['cuda', 'cpu'], default='cuda', help='Select the device: cuda or cpu, gpu memory larger than 14G is recommended')
parser.add_argument('--dense_captioner_device', choices=['cuda', 'cpu'], default='cuda', help='Select the device: cuda or cpu, < 6G GPU is not recommended>')
parser.add_argument('--audio_translator', default='large')
parser.add_argument('--audio_translator_device', choices=['cuda', 'cpu'], default='cuda')
parser.add_argument('--gpt_version', choices=['gpt-3.5-turbo'], default='gpt-3.5-turbo')

args = parser.parse_args()

vlogger = Vlogger4chat(args)

def get_empty_state():
    return {"total_tokens": 0, "messages": []}


def submit_message(prompt, state):
    history = state['messages']

    if not prompt:
        return gr.update(value=''), [(history[i]['content'], history[i+1]['content']) for i in range(0, len(history)-1, 2)], state

    prompt_msg = { "role": "user", "content": prompt }
    
    try:
        history.append(prompt_msg)
        answer = vlogger.chat2video(prompt)
        history.append({"role": "system", "content": answer}) 

    
    except Exception as e:
        history.append(prompt_msg)
        history.append({
            "role": "system",
            "content": f"Error: {e}"
        })

    chat_messages = [(history[i]['content'], history[i+1]['content']) for i in range(0, len(history)-1, 2)]
    return '', chat_messages, state


def submit_message_debate(prompt, state):
    history = state['messages']
    
    if not prompt:
        return gr.update(value=''), [(history[i]['content'], history[i+1]['content']) for i in range(0, len(history)-1, 2)], state

    prompt_msg = { "role": "user", "content": prompt }
    
    try:
        history.append(prompt_msg)
        
        debate_topic = ""
        while debate_topic == "":
            debate_topic = prompt

        config = json.load(open("./config4all.json", "r"))
        config['debate_topic'] = debate_topic

        debate = Debate(num_players=3, config=config, temperature=0, sleep_time=0)
        answer = debate.run()

        #chat_messages = [(res["debate_topic"]), (res["base_answer"]), (res["debate_answer"]), (res["Reason"])]
        history.append({"role": "system", "content": answer}) 

    
    except Exception as e:
        history.append(prompt_msg)
        history.append({
            "role": "system",
            "content": f"Error: {e}"
        })

    chat_messages = [(history[i]['content'], history[i+1]['content']) for i in range(0, len(history)-1, 2)]
    return '', chat_messages, state


def clear_conversation():
    vlogger.clean_history()
    return gr.update(value=None, visible=True), gr.update(value=None, visible=True), gr.update(value=None, interactive=True), None, gr.update(value=None, visible=True), get_empty_state()


# download video from any online URL 
def subvid_fn(vid):
    print(vid)
    save_path = download_video(vid)
    return gr.update(value=save_path)


# 本地上传，适用于Running on local URL:  http://127.0.0.1:6006
def uploaded_video(video_file):
    UPLOAD_FOLDER = "./"   
    if not os.path.exists(UPLOAD_FOLDER):    
        os.mkdir(UPLOAD_FOLDER)    
    shutil.copy(video_file, UPLOAD_FOLDER)    
    gr.Info("File Uploaded!!!")   
    save_path = os.path.join(UPLOAD_FOLDER, os.path.basename(video_file))
    return gr.update(value=save_path)

    
def vlog_fn(vid_path):
    print(vid_path)
    if vid_path is None:
        log_text = "====== Please choose existing video from the library or provide video URL 🤔====="
    else:
        log_list = vlogger.video2log(vid_path)
        log_text = "\n".join(log_list)
    return gr.update(value=log_text, visible=True)

# 初始化一个空的答案记录字典
answers = {}

# 定义处理用户选择的函数
def submit_answers_pretest(question1, question2, question3, question4, question5, question6, question7, question8, question9, question10):
    answers['Question 1'] = question1
    answers['Question 2'] = question2
    answers['Question 3'] = question3
    answers['Question 4'] = question4
    answers['Question 5'] = question5
    answers['Question 6'] = question6
    answers['Question 7'] = question7
    answers['Question 8'] = question8
    answers['Question 9'] = question9
    answers['Question 10'] = question10
    
    # 可以将结果保存到文件
    with open('answers4pretest.txt', 'a') as f:
        f.write(f"Question 1: {question1}\n")
        f.write(f"Question 2: {question2}\n")
        f.write(f"Question 3: {question3}\n")
        f.write(f"Question 4: {question4}\n")
        f.write(f"Question 5: {question5}\n")
        f.write(f"Question 6: {question6}\n")
        f.write(f"Question 7: {question7}\n")
        f.write(f"Question 8: {question8}\n")
        f.write(f"Question 9: {question9}\n")
        f.write(f"Question 10: {question10}\n\n")
    
    # 返回一个确认消息
    return "谢谢你提交答案!"

def submit_answers_posttest(question1, question2, question3, question4, question5, question6, question7, question8, question9, question10):
    answers['Question 1'] = question1
    answers['Question 2'] = question2
    answers['Question 3'] = question3
    answers['Question 4'] = question4
    answers['Question 5'] = question5
    answers['Question 6'] = question6
    answers['Question 7'] = question7
    answers['Question 8'] = question8
    answers['Question 9'] = question9
    answers['Question 10'] = question10
    
    # 可以将结果保存到文件
    with open('answers4posttest.txt', 'a') as f:
        f.write(f"Question 1: {question1}\n")
        f.write(f"Question 2: {question2}\n")
        f.write(f"Question 3: {question3}\n")
        f.write(f"Question 4: {question4}\n")
        f.write(f"Question 5: {question5}\n")
        f.write(f"Question 6: {question6}\n")
        f.write(f"Question 7: {question7}\n")
        f.write(f"Question 8: {question8}\n")
        f.write(f"Question 9: {question9}\n")
        f.write(f"Question 10: {question10}\n\n")
    
    # 返回一个确认消息
    return "谢谢你提交答案!"

css = """
      #col-container {max-width: 80%; margin-left: auto; margin-right: auto;}
      #video_inp {min-height: 100px}
      #chatbox {min-height: 100px;}
      #header {text-align: center;}
      #hint {font-size: 2.0em; padding: 0.5em; margin: 0;}
      .message { font-size: 1.2em; }
      """

with gr.Blocks(css=css) as demo:
    
    with gr.Tabs():
        # 第一个标签页
        with gr.TabItem("第一步(观看前测试)"):
            gr.Markdown("""## 在观看视频前，我们先进行一个简单的测试: 由和视频内容相关的六道题目组成
                        请二选一：访问 [正则化](https://www.wjx.cn/vm/PBZtmpe.aspx#) 进行答题，
                                或者[梯度的数值逼近](https://kaoshi.wjx.top/vm/YDQNWYC.aspx#) 进行答题。""",
                                elem_id="header")
        # 第二个标签页
        with gr.TabItem("第二步(VLog使用)"):
            state = gr.State(get_empty_state())

            with gr.Column(elem_id="col-container"):
                gr.Markdown("""## 🎞️ 视频Chat: 
                            Powered by CLIP, BLIP2, GRIT, RAM++, PaddleOCR, Whisper, Custom LLMs and LangChain""",
                            elem_id="header")

                with gr.Row():
                    with gr.Column():
                        video_inp = gr.Video(label="video_input")

                        gr.Markdown("Step 1: 请在下方选取视频(Select videos below)", elem_id="hint")
                        examples = gr.Examples(
                            examples=[
                                ["./BV11H4y1F7uH-P50.mp4"],
                                ["./BV11H4y1F7uH-P58.mp4"],
                            ],
                            inputs=[video_inp],
                        )
                        
                        #with gr.Row():
                        #    video_id = gr.Textbox(value="", placeholder="Download video url", show_label=False)
                        #    vidsub_btn = gr.Button("上传网站视频")

                        chatbot = gr.Chatbot(elem_id="chatbox")
                        input_message = gr.Textbox(show_label=False, placeholder="输入提问内容并按回车(Input your question)", visible=True)
                        btn_submit = gr.Button("提问视频内容(Submit)")
                        
                        gr.Markdown("如果对上面的回答不满意，请在下方输入需要辩论的问题, *e.g.* *方差越小越好? 正则化必不可缺? 梯度一定存在?*", elem_id="hint")
                        #chatbot_debate = gr.Chatbot(elem_id="chatbox")
                        input_message_debate = gr.Textbox(show_label=False, placeholder="输入辩论主题并按回车(Input your debate topic)", visible=True)
                        btn_submit_debate = gr.Button("发起问题辩论(Submit)")
                        
                        btn_clear_conversation = gr.Button("🔃 开始新的对话(Start new conversation)")

                    with gr.Column():
                        vlog_btn = gr.Button("Step 2: 查看视频日志(Pleas wait around 20 seconds)")
                        vlog_outp = gr.Textbox(label="Document output", lines=70)
                        total_tokens_str = gr.Markdown(elem_id="total_tokens_str")

                

            gr.HTML('''<br><br><br><center>You can duplicate this Space to skip the queue:<a href="https://huggingface.co./spaces/anzorq/chatgpt-demo?duplicate=true"><img src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a><br></center>''')

            btn_submit.click(submit_message, [input_message, state], [input_message, chatbot])
            input_message.submit(submit_message, [input_message, state], [input_message, chatbot])
            btn_submit_debate.click(submit_message_debate, [input_message_debate, state], [input_message_debate, chatbot])
            input_message_debate.submit(submit_message_debate, [input_message_debate, state], [input_message_debate, chatbot])
            btn_clear_conversation.click(clear_conversation, [], [input_message, input_message_debate, video_inp, chatbot, vlog_outp, state])
            vlog_btn.click(vlog_fn, [video_inp], [vlog_outp])
            #vidsub_btn.click(subvid_fn, [video_id], [video_inp])
            
        # 第三个标签页
        with gr.TabItem("第三步(观后测试)"):
            gr.Markdown("""## 在观看视频后，我们再进行一个简单的测试: 也是由和视频内容相关的六道题目组成
                        请二选一：访问 [正则化](https://www.wjx.cn/vm/PBZtmpe.aspx#) 进行答题，
                                或者[梯度的数值逼近](https://kaoshi.wjx.top/vm/YDQNWYC.aspx#) 进行答题。""",
                                elem_id="header")
        

    demo.load(queue=False)


demo.queue()

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