前些天发现了一个巨牛的人工智能学习网站，通俗易懂，风趣幽默，忍不住分享一下给大家：点击跳转
目录
一，合并ts文件
1，背景
当你从网站下载了一集动漫，然后发现是一堆ts文件，虽然可以打开，但是每个都是10秒左右，很不方便。
这时，可以用python合并ts文件。
2，安装配置ffmpeg
官网下载ffmpeg-2023-08-07-git-d295b6b693-full_build
放到一个没有空格的路径中：
再把bin目录配置到环境变量。
官网下载ffmpeg
下载之后复制到site-packages中：
3，实现
代码：
def merge_file(path, save_path):
file_names = os.listdir(path)
if 'file_list.txt' in file_names:
os.remove(path+'file_list.txt')
f = open(path+'file_list.txt', 'w+')
for one in file_names:
f.write("file '" + one + "'\n")
f.close()
ffmpeg_bin_dic = 'C:/ffmpeg-2023-08-07-git-d295b6b693-full_build/bin/'
os.system(ffmpeg_bin_dic+'ffmpeg -f concat -safe 0 -i '+path+'file_list.txt'+' -c '+ ' copy ' +save_path)注意，需要看一下file_list.txt中的各个ts顺序，和预期是否一致。
如果不一致，调整原有ts的命名，或者调整代码，直到file_list.txt中的各个ts顺序，和预期一致，才能得到正确的mp4文件。
4，缺点
貌似合并出来的视频经常有问题。。。
二，下载m3u8视频
1，背景
当你有一个m3u8视频的地址，那就可以解析m3u8，获取其中的ts列表，多线程下载所有ts文件。
全部下完之后，用ffmpeg合并成mp4
2，实现V1
Downloader把ts列表做循环拆分，近似均分给每个线程，同时下载1个m3u8视频。
download_m3u8_video输入m3u8视频的地址，解析ts列表，调用Downloader并控制多线程。
import requests
import os
import threading
tnum = 64
class Downloader(threading.Thread):
def __init__(self, id, url, ts_list, file_path):
threading.Thread.__init__(self)
self.id = id
self.url = url
self.ts_list = ts_list
self.file_path = file_path
def run(self):
for i in range(self.id, len(self.ts_list), tnum):
ts_url = self.ts_list[i]
r = requests.get(ts_url, stream=True)
if r.status_code == 200:
with open(self.file_path + f'_{i}.ts', 'wb') as f:
for chunk in r.iter_content(chunk_size=1024):
if chunk:
f.write(chunk)
def download_m3u8_video(url, file_path):
r = requests.get(url)
if r.status_code != 200:
print('m3u8视频下载链接无效')
return False
m3u8_list = r.text.split('\n')
m3u8_list = [i for i in m3u8_list if i and i[0] != '#']
ts_list = []
for ts_url in m3u8_list:
ts_url = url.rsplit('/', 1)[0] + '/' + ts_url
ts_list.append(ts_url)
f = open(file_path+'file_list.txt', 'w+')
for i in range(len(ts_list)):
f.write('file \'' + f'_{i}.ts' + "'\n")
f.close()
threads = []
for i in range(0, tnum):
thread = Downloader(i, url, ts_list, file_path.rsplit('.', 1)[0])
thread.start()
threads.append(thread)
for thread in threads:
thread.join()
return not endflag
def merge_file(path, save_path):
ffmpeg_bin_dic = 'C:/ffmpeg-2023-08-07-git-d295b6b693-full_build/bin/'
os.system(ffmpeg_bin_dic+'ffmpeg -f concat -safe 0 -i '+path+'file_list.txt'+' -c '+ ' copy ' +save_path)
def down(url):
name = url[-41:-5]
print(name)
ts_file_path = 'D:/v/' + name+'/'
os.makedirs(ts_file_path)
mp4_file_path = 'D:/v2/'+name+'.mp4'
global endflag
endflag = False
if download_m3u8_video(url, ts_file_path):
print('m3u8视频下载完成')
merge_file(ts_file_path, mp4_file_path)
else :
print('error!!!!')
print(url)
url_list = [
'https://sth.com/videos/202305091/d63b006e-c6fb-4997-8d43-7ebd086e9c75.m3u8',
'https://sth.com/videos/202305081/2c9fa41b-e25b-4371-908f-d246628d7bed.m3u8',
]
for url in url_list:
down(url)3，实现V2
import requests
import os
import threading
tnum = 64
def download_m3u8_video(url, file_path):
r = requests.get(url)
if r.status_code != 200:
return False
m3u8_list = r.text.split('\n')
m3u8_list = [i for i in m3u8_list if i and i[0] != '#']
ts_list = []
for ts_url in m3u8_list:
ts_url = url.rsplit('/', 1)[0] + '/' + ts_url
ts_list.append(ts_url)
f = open(file_path + 'file_list.txt', 'w+')
for i in range(len(ts_list)):
f.write('file \'' + f'_{i}.ts' + "'\n")
f.close()
for i in range(0, len(ts_list)):
ts_url = ts_list[i]
r = requests.get(ts_url, stream=True)
if r.status_code == 200:
with open(file_path + f'_{i}.ts', 'wb') as f:
for chunk in r.iter_content(chunk_size=1024):
if chunk:
f.write(chunk)
return True
def merge_file(path, save_path):
ffmpeg_bin_dic = 'C:/ffmpeg-2023-08-07-git-d295b6b693-full_build/bin/'
os.system(ffmpeg_bin_dic+'ffmpeg -f concat -safe 0 -i '+path+'file_list.txt'+' -c '+ ' copy ' +save_path)
def down(url):
name = url[-41:-5]
print(name)
ts_file_path = 'D:/v/' + name+'/'
os.makedirs(ts_file_path)
mp4_file_path = 'D:/v2/'+name+'.mp4'
ret = download_m3u8_video(url, ts_file_path)
print(name, end=' ')
if ret:
print('m3u8视频下载完成')
merge_file(ts_file_path, mp4_file_path)
else :
print('error!!!!')
class down_list(threading.Thread):
def __init__(self, url_list, id):
threading.Thread.__init__(self)
self.id = id
self.url_list = url_list
def run(self):
for i in range(self.id, len(self.url_list), tnum):
url = self.url_list[i]
down(url)
def down_all(url_list):
threads=[]
for i in range(0, tnum):
thread = down_list(url_list,i)
thread.start()
threads.append(thread)
for thread in threads:
thread.join()
url_list = [
]
down_all(url_list)