B站电脑客户端缓存视频合成

假日无事，手撸几行代码实现B站缓存数据视频合成，主要原理在：

1、查找m4s文件，二进制编辑，去除前9个0，并本地保存

2、调用ffmpeg进行音视频合成，需要注意中文路径问题（下载中文版的ffmpeg，放到系统path中）

代码如下：

#!/usr/bin/python
# -*- coding: utf-8 -*-
import sys
import os
import glob
import re
import time
import json
import io
import argparse
import concurrent.futures
import subprocess
import ffmpeg

def parse_args():
    parser = argparse.ArgumentParser(description='video and audio m4s to mp4')
    parser.add_argument('--workdir', type=str, default='D:\\bilibili\\', help='the dir to work ')
    parser.add_argument('--num_threads', type=int, default=8, help='the num of threads to combine video and audio file to mp4')
    args = parser.parse_args()
    return args


def remove9zero(fn):
    with open(fn, 'rb+') as f:
        data = f.read()
        f.seek(0)
        f.write(data[9:])
        f.truncate()


def renamefile(oldfn, newfn):
    try:
        os.rename(oldfn, newfn)
        print("rename complete！")
    except FileNotFoundError:
        print("file not found！")
    except Exception as e:
        print("error：", str(e))


def getfileInfo(folder):
    result = {}
    videoinfo = folder +'\\.videoInfo'
    with open(videoinfo, 'r', encoding='UTF-8') as f:
        data = json.load(f)
        #print(data['title'])
        result['title'] = data['title']
        result['folder_title'] = data['groupTitle']
    return result


def validateTitle(title):
    rstr = r"[\/\\\:\*\?\"\<\>\|]"
    new_title = re.sub(rstr, "_", str(title))
    return new_title


def ffmpeg2mp4(f1, f2, fn, folder):
    str1 = f'ffmpeg.exe -i {f1} -i {f2} -c copy {args.workdir} video\\{folder} \\{fn} .mp4'
    print(f'ffmpeg.exe -i {f1} -i {f2} -c copy {args.workdir} video\\{folder} \\{fn}.mp4')
    os.system(f'ffmpeg.exe -i {f1} -i {f2} -c copy {args.workdir} video\\{folder} \\{fn}.mp4')

def combinefile(argslist, args):
    '''with concurrent.futures.ThreadPoolExecutor(
            max_workers=int(args.num_threads)
    ) as executor:
        for value in argslist:
            f1 = value['f1']
            f2 = value['f2']
            fn = value['fn']
            folder = value['folder']
            executor.submit(ffmpeg2mp4, f1, f2, fn, folder)
        executor.shutdown()'''
    for value in argslist:
        f1 = value['f1']
        f2 = value['f2']
        fn = value['fn'].replace(' ', '')
        folder = value['folder'].replace(' ', '')
        #os.system(f'ffmpeg.exe -i {f1} -i {f2} -c copy {args.workdir}video\\{folder}\\{fn}.mp4')
        #os.system(f'ffmpeg.exe -i {f1} -i {f2} -c copy {args.workdir}video\\temp.mp4')
        #subprocess.run(f'ffmpeg.exe -i {f1} -i {f2} -y -c copy {args.workdir}video\\temp.mp4', encoding="utf-8")
        #subprocess.run(f'move {args.workdir}video\\temp.mp4 {args.workdir}video\\{folder}\\{fn}.mp4', encoding="utf-8")
        subprocess.run(f'ffmpeg.exe -i {f1} -i {f2} -y -c copy {args.workdir}video\\{folder}\\{fn}.mp4')
        print(f'ffmpeg.exe -i {f1} -i {f2} -y -c copy {args.workdir}video\\{folder}\\{fn}.mp4')

def main():
    args = parse_args()
    # 查找.videoInfo文件到列表，并将列表目录存入集合
    files_videoinfo = glob.glob(args.workdir + "download\\*\\.videoInfo")
    fv_set = set()
    for fi in files_videoinfo:
        fv_set.add(os.path.dirname(fi))

    # 查找偶数个.m4s文件的目录，并存入folder_list
    folder_list = []
    folder_num = 0
    for item in fv_set:
        file_dir = glob.glob(item + "\\*.m4s")
        #file_dir = glob.glob(item + "\\*.mp4")
        if len(file_dir) % 2 != 0:
            print(item + 'have not ' + str(len(file_dir)) +'.m4s files')
        elif len(file_dir) == 2:
            folder_num = folder_num + 1
            folder_list.append(item)
        else:
            print(item)
    print('total folder num is ' + str(folder_num))

    # 对folder_list进行遍历处理，移除前9个0，重命名为mp4，获取对应的中文名称
    argslist = []
    for folder in folder_list:
        files = glob.glob(folder + "\\*.m4s")
        #files = glob.glob(folder + "\\*.mp4")
        for file in files:
            remove9zero(file)
            renamefile(file, file[:-3] + 'mp4')

        dict_item = {}
        finfo = getfileInfo(folder)
        grouptitle = validateTitle(finfo['folder_title']).replace(' ', '')
        title = validateTitle(finfo['title']).replace(' ', '')
        dict_item.clear()
        dict_item['f1'] = files[0][:-3] + 'mp4'
        dict_item['f2'] = files[1][:-3] + 'mp4'
        dict_item['fn'] = title
        dict_item['folder'] = grouptitle
        argslist.append(dict_item)

    # 查找folder的数量，并创建对应的文件夹
    folder_list = set()
    for argitem in argslist:
        folder_list.add(argitem['folder'])
    print('folder num is ' + str(len(folder_list)))
    for folder_name in folder_list:
        path = args.workdir + 'video\\' + folder_name
        if not os.path.exists(path):
            os.mkdir(path)
    # 多线程进行视频音频合并
    combinefile(argslist, args)


if __name__ == '__main__':
    main()