python获取问津

import os
import requests
import json
import re
from bs4 import BeautifulSoup
from qiniu import Auth, put_file


class BilibiliVideoAudio:
    def __init__(self, bid, download_path='download'):
        self.bid = bid
        self.headers = {
            "referer": "https://www.bilibili.com",
            "origin": "https://www.bilibili.com",
            'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.129 Safari/537.36',
            'Accept-Encoding': 'identity'
        }
        self.download_path = download_path

        # 创建下载路径
        if not os.path.exists(download_path):
            os.makedirs(download_path)

    def get_video_audio(self):
        # 构造视频链接并发送请求获取页面内容
        url = f'https://www.bilibili.com/video/{self.bid}?vd_source=6659c978e0a0ff11c82fae0ae309aabe'
        content = requests.get(url, headers=self.headers).content.decode('utf-8')
        soup = BeautifulSoup(content, 'html.parser')

        # 获取视频标题
        meta_tag = soup.head.find('meta', attrs={'name': 'title'})
        title = meta_tag['content']

        # 获取视频和音频链接
        pattern = r'window\.__playinfo__=({.*?})\s*</script>'
        json_data = re.findall(pattern, content)[0]
        data = json.loads(json_data)

        video_url = data['data']['dash']['video'][0]['base_url']
        audio_url = data['data']['dash']['audio'][0]['base_url']

        return {
            'title': title,
            'video_url': video_url,
            'audio_url': audio_url
        }

    def download_video_audio(self, url, filename):
        # 对文件名进行清理，去除不合规字符
        filename = self.sanitize_filename(filename)
        try:
            # 发送请求下载视频或音频文件
            resp = requests.get(url, headers=self.headers).content
            download_path = os.path.join(self.download_path, filename)
            with open(download_path, mode='wb') as file:
                file.write(resp)
            print("{:*^30}".format(f"下载完成：{filename}"))

            return download_path  # 返回下载文件的本地路径
        except Exception as e:
            print(e)

    def sanitize_filename(self, filename):
        # 定义不合规字符的正则表达式
        invalid_chars_regex = r'[\"*<>?\\|/:,]'

        # 替换不合规字符为空格
        sanitized_filename = re.sub(invalid_chars_regex, ' ', filename)

        return sanitized_filename


def upload_to_qiniu(access_key, secret_key, bucket_name, local_file_path, key=None):
    q = Auth(access_key, secret_key)
    token = q.upload_token(bucket_name, key, 3600)
    ret, info = put_file(token, key, local_file_path)
    if info.status_code == 200:
        print("文件上传成功！")
    else:
        print("文件上传失败：", info.error)


def main():
    bids = ["BV1bL4y1E71a"]  # 视频的bid，可以修改为其他视频的bid
    for bid in bids:
        bilibili = BilibiliVideoAudio(bid)
        video_audio_info = bilibili.get_video_audio()

        title = video_audio_info['title']
        video_url = video_audio_info['video_url']
        audio_url = video_audio_info['audio_url']

        video_path = bilibili.download_video_audio(video_url, f"{title}.mp4")  # 下载视频
        audio_path = bilibili.download_video_audio(audio_url, f"{title}.mp3")  # 下载音频

        # 上传视频和音频到七牛云
        upload_to_qiniu(access_key='lObvgTkNUTdmLkeRH-fSuc3jeHEfdAgtxXzzvIbW',
                        secret_key='YjnzJuCH_6-qWb8aQ5nGTIqsVDFywRN_tgdrkOF_',
                        bucket_name='shanjupayss',
                        local_file_path=video_path,
                        key=f"{title}.mp4")
        upload_to_qiniu(access_key='lObvgTkNUTdmLkeRH-fSuc3jeHEfdAgtxXzzvIbW',
                        secret_key='YjnzJuCH_6-qWb8aQ5nGTIqsVDFywRN_tgdrkOF_',
                        bucket_name='shanjupayss',
                        local_file_path=audio_path,
                        key=f"{title}.mp3")


if __name__ == "__main__":
    main()