可以發現B站視訊是分片獲取的,並且先發起options預請求,在發起get請求獲取視訊資料,視訊和音訊檔案url路徑也有不同,下面定位兩個url連結位置

通過請求獲取攜帶的引數名定位視訊url連結在頁面原始碼中存在的位置

對頁面發起請求,使用etree封裝獲取到的頁面文件,使用xpath定位字串位置,再通過json模組將字串轉化為json字串,獲取視訊url地址和音訊url地址,程式碼如下

• def GetBiliVideo(url,session=requests.session()):
     res = requests.get(url=url,headers = headers)
   
     html = etree.HTML(res.content)
     videoInfo = html.xpath('/html/head/script[3]/text()') # 解析出真實地址存在的位置
     # videoJson = json.loads(str(videoInfo[0].split('=')[1]))
     videoJson = json.loads(videoInfo[0].split('=',1)[1]) # 切出符合轉化要求的字串並轉化為json字串
     videoUrl = videoJson['data']['dash']['video'][0]['baseUrl']  #視訊地址
   
     audioUrl = videoJson['data']['dash']['audio'][0]['baseUrl'] # 音訊地址
     BiliDown(videoUrl,audioUrl) # 呼叫下載函式

發起請求,並將獲取的資料寫成檔案,

• def BiliDown(url1,url2,session = requests.session()):
     headers.update({'Referer':url})
     session.options(url=url1, headers=headers)  # 先用options方法檢視伺服器資源
     res = session.get(url=url1, headers=headers, ) # 然後get獲取視訊二進位制檔案
   
     session.options(url=url2, headers=headers)  # 音訊的
     res2 = session.get(url=url2, headers=headers, )
     with open(r'E:\ES\video\bili.mp4', 'wb') as fp:  # 存視訊
       fp.write(res.content)
  ​
     with open(r'E:\ES\video\bili1.mp4', 'wb') as fp1:  # 存音訊
       fp1.write(res2.content)                  
  ​

可以獲得視訊和音訊檔案,通過ffmpeg將視訊和音訊合成

• def combine_audio(video_file, audiio_file, out_file):
     try:
       cmd =r'E:\ffmpeg\ffmpeg-4.2.2-win64-static\bin/ffmpeg -i '+video_file+' -i '+audiio_file+' -acodec copy '+out_file
       print(cmd)
       subprocess.call(cmd, shell=True)  # "Muxing Done
       print('done!!!')
       return True
     except Exception:
       return False

• 速度有點慢,14分鐘音訊視訊大概花了7分鐘來合成.

整體程式碼如下,尚未完善,只是做了單個視訊爬取

• import requests
  import json
  from lxml import etree
  import os
  import subprocess
  ​
  url = 'https://www.bilibili.com/video/BV1Mi4y1G7ZH'  # 視訊頁面地址
  ​
  headers = {
  'Referer': 'https://www.bilibili.com/',
  'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/73.0.3683.75 Safari/537.36'
  }
  ​
  def GetBiliVideo(url,session=requests.session()):
     res = requests.get(url=url,headers = headers)
     html = etree.HTML(res.content)
     videoInfo = html.xpath('/html/head/script[3]/text()') # 解析出真實地址存在的位置
     # videoJson = json.loads(str(videoInfo[0].split('=')[1]))
     videoJson = json.loads(videoInfo[0].split('=',1)[1]) # 切出符合轉化要求的字串並轉化為json字串
     videoUrl = videoJson['data']['dash']['video'][0]['baseUrl']  #視訊地址
     audioUrl = videoJson['data']['dash']['audio'][0]['baseUrl'] # 音訊地址
     BiliDown(videoUrl,audioUrl) # 呼叫下載函式
  def BiliDown(url1,url2,session = requests.session()):
     headers.update({'Referer':url})
     session.options(url=url1, headers=headers)  # 先用options方法檢視伺服器資源
     res = session.get(url=url1, headers=headers, ) # 然後get獲取視訊二進位制檔案
     session.options(url=url2, headers=headers)  # 音訊的
     res2 = session.get(url=url2, headers=headers, )
     with open(r'E:\ES\video\bili.mp4', 'wb') as fp:  # 存視訊
       fp.write(res.content)
  ​
     with open(r'E:\ES\video\bili1.mp4', 'wb') as fp1:  # 存音訊
       fp1.write(res2.content)
  ​
     # path都需要使用全路徑
     combine_audio(r'E:\ES\video\bili.mp4',r'E:\ES\video\bili1.mp4',r'E:\ES\video\bili2.mp4')
  ​
  # 組合音訊和視訊 （自己加的）
  def combine_audio(video_file, audiio_file, out_file):
     try:
       cmd =r'E:\ffmpeg\ffmpeg-4.2.2-win64-static\bin/ffmpeg -i '+video_file+' -i '+audiio_file+' -acodec copy '+out_file
       print(cmd)
       subprocess.call(cmd, shell=True)  # "Muxing Done
       print('done!!!')
       return True
     except Exception:
       return False
  GetBiliVideo(url)