app.py

import gradio as gr
import pandas as pd
import requests
from bs4 import BeautifulSoup
from docx import Document
import os
from openai import OpenAI
import json

from youtube_transcript_api import YouTubeTranscriptApi
from youtube_transcript_api._errors import NoTranscriptFound


from moviepy.editor import VideoFileClip
from pytube import YouTube
import os

from google.oauth2 import service_account
from googleapiclient.discovery import build
from googleapiclient.http import MediaFileUpload
from googleapiclient.http import MediaIoBaseDownload
from googleapiclient.http import MediaIoBaseUpload

import io


from urllib.parse import urlparse, parse_qs


# 假设您的环境变量或Secret的名称是GOOGLE_APPLICATION_CREDENTIALS_JSON
# credentials_json_string = os.getenv("GOOGLE_APPLICATION_CREDENTIALS_JSON")
# credentials_dict = json.loads(credentials_json_string)
# SCOPES = ['https://www.googleapis.com/auth/drive']
# credentials = service_account.Credentials.from_service_account_info(
#         credentials_dict, scopes=SCOPES)
# service = build('drive', 'v3', credentials=credentials)
# # 列出 Google Drive 上的前10個文件
# results = service.files().list(pageSize=10, fields="nextPageToken, files(id, name)").execute()
# items = results.get('files', [])

# if not items:
#     print('No files found.')
# else:
#     print("=====Google Drive 上的前10個文件=====")
#     print('Files:')
#     for item in items:
#         print(u'{0} ({1})'.format(item['name'], item['id']))



OUTPUT_PATH = 'videos'
TRANSCRIPTS = []
CURRENT_INDEX = 0

OPEN_AI_KEY = os.getenv("OPEN_AI_KEY")
client = OpenAI(api_key=OPEN_AI_KEY)


# # ====drive====初始化Google Drive服务
def init_drive_service():
    credentials_json_string = os.getenv("GOOGLE_APPLICATION_CREDENTIALS_JSON")
    credentials_dict = json.loads(credentials_json_string)
    SCOPES = ['https://www.googleapis.com/auth/drive']
    credentials = service_account.Credentials.from_service_account_info(
            credentials_dict, scopes=SCOPES)
    service = build('drive', 'v3', credentials=credentials)
    return service

def create_folder_if_not_exists(service, folder_name, parent_id):
    print("检查是否存在特定名称的文件夹，如果不存在则创建")
    query = f"mimeType='application/vnd.google-apps.folder' and name='{folder_name}' and '{parent_id}' in parents and trashed=false"
    response = service.files().list(q=query, spaces='drive', fields="files(id, name)").execute()
    folders = response.get('files', [])
    if not folders:
        # 文件夹不存在，创建新文件夹
        file_metadata = {
            'name': folder_name,
            'mimeType': 'application/vnd.google-apps.folder',
            'parents': [parent_id]
        }
        folder = service.files().create(body=file_metadata, fields='id').execute()
        return folder.get('id')
    else:
        # 文件夹已存在
        return folders[0]['id']

# 检查Google Drive上是否存在文件
def check_file_exists(service, folder_name, file_name):
    query = f"name = '{file_name}' and '{folder_name}' in parents and trashed = false"
    response = service.files().list(q=query).execute()
    files = response.get('files', [])
    return len(files) > 0, files[0]['id'] if files else None


def upload_to_drive(service, file_name, folder_id, content):
    print("上传文本内容到Google Drive指定的文件夹中")
    # 如果您的内容是字符串（文本），请使用io.StringIO
    # 对于二进制内容，请使用io.BytesIO
    file_metadata = {'name': file_name, 'parents': [folder_id]}
    # 这里我们假定content是文本，因此使用io.StringIO
    media = MediaFileUpload(io.StringIO(content), mimetype='text/plain')
    
    service.files().create(body=file_metadata, media_body=media, fields='id').execute()

def upload_content_directly(service, file_name, folder_id, content):
    """
    直接将内容上传到Google Drive中的新文件。
    """
    file_metadata = {'name': file_name, 'parents': [folder_id]}
    # 使用io.StringIO为文本内容创建一个内存中的文件对象
    fh = io.BytesIO(content.encode('utf-8'))
    media = MediaIoBaseUpload(fh, mimetype='text/plain', resumable=True)
    
    # 执行上传
    file = service.files().create(body=file_metadata, media_body=media, fields='id').execute()
    return file.get('id')

def download_file_as_string(service, file_id):
    """
    从Google Drive下载文件并将其作为字符串返回。
    """
    request = service.files().get_media(fileId=file_id)
    fh = io.BytesIO()
    downloader = MediaIoBaseDownload(fh, request)
    done = False
    while done is False:
        status, done = downloader.next_chunk()
    fh.seek(0)
    content = fh.read().decode('utf-8')
    return content

def upload_img_directly(service, file_name, folder_id, file_path):
    file_metadata = {'name': file_name, 'parents': [folder_id]}
    media = MediaFileUpload(file_path, mimetype='image/jpeg')
    file = service.files().create(body=file_metadata, media_body=media, fields='id').execute()
    return file.get('id')  # 返回文件ID

def set_public_permission(service, file_id):
    service.permissions().create(
        fileId=file_id,
        body={"type": "anyone", "role": "reader"},
        fields='id',
    ).execute()

def update_file_on_drive(service, file_id, file_content):
    """
    更新Google Drive上的文件内容。
    
    参数:
    - service: Google Drive API服务实例。
    - file_id: 要更新的文件的ID。
    - file_content: 新的文件内容，字符串格式。
    """
    # 将新的文件内容转换为字节流
    fh = io.BytesIO(file_content.encode('utf-8'))
    media = MediaIoBaseUpload(fh, mimetype='application/json', resumable=True)
    
    # 更新文件
    updated_file = service.files().update(
        fileId=file_id,
        media_body=media
    ).execute()
    
    print(f"文件已更新，文件ID: {updated_file['id']}")


# ====drive====

def process_file(file):
    # 读取文件
    if file.name.endswith('.csv'):
        df = pd.read_csv(file)
        text = df_to_text(df)
    elif file.name.endswith('.xlsx'):
        df = pd.read_excel(file)
        text = df_to_text(df)
    elif file.name.endswith('.docx'):
        text = docx_to_text(file)
    else:
        raise ValueError("Unsupported file type")

    df_string = df.to_string()
    # 宜蘭：移除@XX@符号 to |
    df_string = df_string.replace("@XX@", "|")

    # 根据上传的文件内容生成问题
    questions = generate_questions(df_string)
    df_summarise = generate_df_summarise(df_string)

    # 返回按钮文本和 DataFrame 字符串
    return questions[0] if len(questions) > 0 else "", \
           questions[1] if len(questions) > 1 else "", \
           questions[2] if len(questions) > 2 else "", \
           df_summarise, \
           df_string

def df_to_text(df):
    # 将 DataFrame 转换为纯文本
    return df.to_string()

def docx_to_text(file):
    # 将 Word 文档转换为纯文本
    doc = Document(file)
    return "\n".join([para.text for para in doc.paragraphs])

def format_seconds_to_time(seconds):
    """将秒数格式化为 时:分:秒 的形式"""
    hours = int(seconds // 3600)
    minutes = int((seconds % 3600) // 60)
    seconds = int(seconds % 60)
    return f"{hours:02}:{minutes:02}:{seconds:02}"

def extract_youtube_id(url):
    parsed_url = urlparse(url)
    
    if "youtube.com" in parsed_url.netloc:
        # 对于标准链接，视频ID在查询参数'v'中
        query_params = parse_qs(parsed_url.query)
        return query_params.get("v")[0] if "v" in query_params else None
    elif "youtu.be" in parsed_url.netloc:
        # 对于短链接，视频ID是路径的一部分
        return parsed_url.path.lstrip('/')
    else:
        return None

def get_transcript(video_id):
    languages = ['zh-TW', 'zh-Hant', 'en']  # 優先順序列表
    for language in languages:
        try:
            transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=[language])
            return transcript  # 成功獲取字幕，直接返回結果
        except NoTranscriptFound:
            continue  # 當前語言的字幕沒有找到，繼續嘗試下一個語言
    return None  # 所有嘗試都失敗，返回None

def process_transcript_and_screenshots(video_id):
    print("====process_transcript_and_screenshots====")
    service = init_drive_service()
    parent_folder_id = '1GgI4YVs0KckwStVQkLa1NZ8IpaEMurkL'
    folder_id = create_folder_if_not_exists(service, video_id, parent_folder_id)
    file_name = f'{video_id}_transcript.json'

    # 检查逐字稿是否存在
    exists, file_id = check_file_exists(service, folder_id, file_name)
    if not exists:
        # 从YouTube获取逐字稿并上传
        transcript = get_transcript(video_id)
        if transcript:
            print("成功獲取字幕")
        else:
            print("沒有找到字幕")
        transcript_text = json.dumps(transcript, ensure_ascii=False, indent=2)
        file_id = upload_content_directly(service, file_name, folder_id, transcript_text)
        print("逐字稿已上传到Google Drive")
    else:
        # 逐字稿已存在，下载逐字稿内容
        print("逐字稿已存在于Google Drive中")
        transcript_text = download_file_as_string(service, file_id)
        transcript = json.loads(transcript_text)

    # 处理逐字稿中的每个条目，检查并上传截图
    for entry in transcript:
        if 'img_file_id' not in entry:
            screenshot_path = screenshot_youtube_video(video_id, entry['start'])
            img_file_id = upload_img_directly(service, f"{video_id}_{entry['start']}.jpg", folder_id, screenshot_path)
            set_public_permission(service, img_file_id)
            entry['img_file_id'] = img_file_id
            print(f"截图已上传到Google Drive: {img_file_id}")
    
    # 更新逐字稿文件
    updated_transcript_text = json.dumps(transcript, ensure_ascii=False, indent=2)
    update_file_on_drive(service, file_id, updated_transcript_text)
    print("逐字稿已更新，包括截图链接")

    return transcript

def process_youtube_link(link):
    # 使用 YouTube API 获取逐字稿
    # 假设您已经获取了 YouTube 视频的逐字稿并存储在变量 `transcript` 中
    video_id = extract_youtube_id(link)
    download_youtube_video(video_id, output_path=OUTPUT_PATH)
    transcript = process_transcript_and_screenshots(video_id)

    formatted_transcript = []
    screenshot_paths = []
    for entry in transcript:
        start_time = format_seconds_to_time(entry['start'])
        end_time = format_seconds_to_time(entry['start'] + entry['duration'])
        embed_url = get_embedded_youtube_link(video_id, entry['start'])
        img_file_id = entry['img_file_id']
        screenshot_path = f"https://lh3.googleusercontent.com/d/{img_file_id}=s4000"
        line = {
            "start_time": start_time,
            "end_time": end_time,
            "text": entry['text'],
            "embed_url": embed_url,
            "screenshot_path": screenshot_path
        }
        formatted_transcript.append(line)
        screenshot_paths.append(screenshot_path)

    html_content = format_transcript_to_html(formatted_transcript)
    print("=====html_content=====")
    print(html_content)
    print("=====html_content=====")

    # 基于逐字稿生成其他所需的输出
    questions = generate_questions(transcript)
    # 将 DataFrame 转换为纯文本，並分行
    df_string_output = json.dumps(transcript, ensure_ascii=False, indent=2)
    df_summarise = generate_df_summarise(transcript)

    global TRANSCRIPTS
    TRANSCRIPTS = formatted_transcript

    first_image = formatted_transcript[0]['screenshot_path']
    first_text = formatted_transcript[0]['text']

    # 确保返回与 UI 组件预期匹配的输出
    return questions[0] if len(questions) > 0 else "", \
            questions[1] if len(questions) > 1 else "", \
            questions[2] if len(questions) > 2 else "", \
            df_string_output, \
            df_summarise, \
            html_content, \
            first_image, \
            first_text
            

def format_transcript_to_html(formatted_transcript):
    html_content = ""
    for entry in formatted_transcript:
        html_content += f"<h3>{entry['start_time']} - {entry['end_time']}</h3>"
        html_content += f"<p>{entry['text']}</p>"
        html_content += f"<img src='{entry['screenshot_path']}' width='500px' />"
    return html_content



def get_embedded_youtube_link(video_id, start_time):
    embed_url = f"https://www.youtube.com/embed/{video_id}?start={start_time}&autoplay=1"
    return embed_url

def download_youtube_video(youtube_id, output_path=OUTPUT_PATH):
    # Construct the full YouTube URL
    youtube_url = f'https://www.youtube.com/watch?v={youtube_id}'

    # Create the output directory if it doesn't exist
    if not os.path.exists(output_path):
        os.makedirs(output_path)

    # Download the video
    yt = YouTube(youtube_url)
    video_stream = yt.streams.filter(progressive=True, file_extension='mp4').order_by('resolution').desc().first()
    video_stream.download(output_path=output_path, filename=youtube_id+".mp4")

    print(f"Video downloaded successfully: {output_path}/{youtube_id}.mp4")


def screenshot_youtube_video(youtube_id, snapshot_sec):    
    video_path = f'{OUTPUT_PATH}/{youtube_id}.mp4'
    file_name = f"{youtube_id}_{snapshot_sec}.jpg"
    with VideoFileClip(video_path) as video:
        screenshot_path = f'{OUTPUT_PATH}/{file_name}'
        video.save_frame(screenshot_path, snapshot_sec)
    
    return screenshot_path

# def get_screenshot_from_video(video_link, start_time):
#     # 实现从视频中提取帧的逻辑
#     # 由于这需要服务器端处理，你可能需要一种方法来下载视频，
#     # 并使用 ffmpeg 或类似工具提取特定时间点的帧
#     # 这里只是一个示意性的函数实现
#     screenshot_url = f"[逻辑以提取视频 {video_link} 在 {start_time} 秒时的截图]"
#     return screenshot_url

def process_web_link(link):
    # 抓取和解析网页内容
    response = requests.get(link)
    soup = BeautifulSoup(response.content, 'html.parser')
    return soup.get_text()


def generate_df_summarise(df_string):
    # 使用 OpenAI 生成基于上传数据的问题
    sys_content = "你是一個擅長資料分析跟影片教學的老師，user 為學生，請精讀資料文本，自行判斷資料的種類，使用 zh-TW"
    user_content = f"""
        請根據 {df_string}，判斷這份文本
        如果是資料類型，請提估欄位敘述、資料樣態與資料分析，告訴學生這張表的意義，以及可能的結論與對應方式
        
        如果是影片類型，請提估影片內容，告訴學生這部影片的意義，
        小範圍切出不同段落的相對應時間軸的重點摘要，最多不超過五段
        格式為 【start - end】: 摘要，秒數請無條件捨去小數點，並取整數後，根據秒的整數將時間格式改為 hh:mm:ss
        以及可能的結論與結尾小問答提供學生作反思
    """
    messages = [
        {"role": "system", "content": sys_content},
        {"role": "user", "content": user_content}
    ]
    print("=====messages=====")
    print(messages)
    print("=====messages=====")

    request_payload = {
        "model": "gpt-4-1106-preview",
        "messages": messages,
        "max_tokens": 4000,
    }

    response = client.chat.completions.create(**request_payload)
    df_summarise = response.choices[0].message.content.strip()
    print("=====df_summarise=====")
    print(df_summarise)
    print("=====df_summarise=====")

    return df_summarise

def generate_questions(df_string):
    # 使用 OpenAI 生成基于上传数据的问题

    sys_content = "你是一個擅長資料分析跟影片教學的老師，user 為學生，請精讀資料文本，自行判斷資料的種類，並用既有資料為本質猜測用戶可能會問的問題，使用 zh-TW"
    user_content = f"請根據 {df_string} 生成三個問題，並用 JSON 格式返回 questions:[q1, q2, q3]"
    messages = [
        {"role": "system", "content": sys_content},
        {"role": "user", "content": user_content}
    ]
    response_format = { "type": "json_object" }

    print("=====messages=====")
    print(messages)
    print("=====messages=====")


    request_payload = {
        "model": "gpt-4-1106-preview",
        "messages": messages,
        "max_tokens": 4000,
        "response_format": response_format
    }

    response = client.chat.completions.create(**request_payload)
    questions = json.loads(response.choices[0].message.content)["questions"]
    print("=====json_response=====")
    print(questions)
    print("=====json_response=====")

    return questions

def send_question(question, df_string_output, chat_history):
    # 当问题按钮被点击时调用此函数
    return respond(question, df_string_output, chat_history)

def respond(user_message, df_string_output, chat_history):
    print("=== 變數：user_message ===")
    print(user_message)
    print("=== 變數：chat_history ===")
    print(chat_history)

    sys_content = f"""
        你是一個擅長資料分析跟影片教學的老師，user 為學生
        請用 {df_string_output} 為資料文本，自行判斷資料的種類，
        並進行對話，使用 zh-TW

        如果學生問了一些問題你無法判斷，請告訴學生你無法判斷，並建議學生可以問其他問題
        或者你可以問學生一些問題，幫助學生更好的理解資料

        如果學生的問題與資料文本無關，請告訴學生你無法回答超出範圍的問題
    """
    messages = [
        {"role": "system", "content": sys_content},
        {"role": "user", "content": user_message}
    ]

    print("=====messages=====")
    print(messages)
    print("=====messages=====")


    request_payload = {
        "model": "gpt-4-1106-preview",
        "messages": messages,
        "max_tokens": 4000  # 設定一個較大的值，可根據需要調整
    }

    response = client.chat.completions.create(**request_payload)
    print(response)
    
    response_text = response.choices[0].message.content.strip()

    # 更新聊天历史
    new_chat_history = (user_message, response_text)
    if chat_history is None:
        chat_history = [new_chat_history]
    else:
        chat_history.append(new_chat_history)

    # 返回聊天历史和空字符串清空输入框
    return "", chat_history

def update_slide(direction):
    global TRANSCRIPTS
    global CURRENT_INDEX

    print("=== 更新投影片 ===")
    print(f"CURRENT_INDEX: {CURRENT_INDEX}")
    print(f"TRANSCRIPTS: {TRANSCRIPTS}")

    CURRENT_INDEX += direction
    if CURRENT_INDEX < 0:
        CURRENT_INDEX = 0  # 防止索引小于0
    elif CURRENT_INDEX >= len(TRANSCRIPTS):
        CURRENT_INDEX = len(TRANSCRIPTS) - 1  # 防止索引超出范围
    
    # 获取当前条目的文本和截图 URL
    current_transcript = TRANSCRIPTS[CURRENT_INDEX]
    slide_image = current_transcript["screenshot_path"]
    slide_text = current_transcript["text"]
    
    return slide_image, slide_text

def prev_slide():
    return update_slide(-1)

# 包装函数来处理 "下一个" 按钮点击事件
def next_slide():
    return update_slide(1)



with gr.Blocks() as demo:
    with gr.Row():
        with gr.Column():
            file_upload = gr.File(label="Upload your CSV or Word file")
            youtube_link = gr.Textbox(label="Enter YouTube Link")
            web_link = gr.Textbox(label="Enter Web Page Link")
            chatbot = gr.Chatbot()
            msg = gr.Textbox(label="Message")
            send_button = gr.Button("Send")

        with gr.Column():
            with gr.Tab("截圖與逐字稿"):
                transcript_html = gr.HTML(label="YouTube Transcript and Video") 
            with gr.Tab("投影片"):
                slide_image = gr.Image()
                slide_text = gr.Textbox()
                with gr.Row():
                    prev_button = gr.Button("Previous")
                    next_button = gr.Button("Next")
                prev_button.click(fn=prev_slide, inputs=[], outputs=[slide_image, slide_text])
                next_button.click(fn=next_slide, inputs=[], outputs=[slide_image, slide_text])         
            with gr.Tab("資料本文"):
                df_string_output = gr.Textbox(lines=40, label="Data Text")
            with gr.Tab("資料摘要"):
                gr.Markdown("## 這是什麼樣的資料？")
                df_summarise = gr.Textbox(container=True, show_copy_button=True, label="資料本文", lines=40)    
            with gr.Tab("常用問題"):
                gr.Markdown("## 常用問題")
                btn_1 = gr.Button()
                btn_2 = gr.Button()
                btn_3 = gr.Button()

    send_button.click(
        respond, 
        inputs=[msg, df_string_output, chatbot], 
        outputs=[msg, chatbot]
    )
    # 连接按钮点击事件
    btn_1.click(respond, inputs=[btn_1, df_string_output, chatbot], outputs=[msg, chatbot])
    btn_2.click(respond, inputs=[btn_2, df_string_output, chatbot], outputs=[msg, chatbot])
    btn_3.click(respond, inputs=[btn_3, df_string_output, chatbot], outputs=[msg, chatbot])


    # file_upload.change(process_file, inputs=file_upload, outputs=df_string_output)
    file_upload.change(process_file, inputs=file_upload, outputs=[btn_1, btn_2, btn_3, df_summarise, df_string_output])

    # 当输入 YouTube 链接时触发
    youtube_link.change(process_youtube_link, inputs=youtube_link, outputs=[btn_1, btn_2, btn_3, df_string_output, df_summarise, transcript_html, slide_image, slide_text])

    # 当输入网页链接时触发
    web_link.change(process_web_link, inputs=web_link, outputs=[btn_1, btn_2, btn_3, df_summarise, df_string_output])


if TRANSCRIPTS:  # 确保列表不为空
    first_screenshot_path, first_text = update_slide(0)
    image.update(value=first_screenshot_path)
    text.update(value=first_text)

demo.launch(allowed_paths=["videos"])