我真的不明白youtube如何提供视频,但我一直在阅读我的能力,似乎旧的方法get_video现在已经过时,不能再使用了因为这个我问的是否有另一个用于收集YouTube视频的pythonic和简单方法。
答案 0 :(得分:15)
youtube-dl
可能会有运气http://rg3.github.com/youtube-dl/documentation.html
我不确定是否有一个好的API,但它是用Python编写的,所以理论上你可以做一些比Popen更好的事情:)
答案 1 :(得分:6)
这是一个下载Youtube视频的快速Python脚本。没有花里胡哨,只需擦除必要的网址,点击generate_204网址,然后将数据流式传输到文件中:
import lxml.html
import re
import sys
import urllib
import urllib2
_RE_G204 = re.compile('"(http:.+.youtube.com.*\/generate_204[^"]+")', re.M)
_RE_URLS = re.compile('"fmt_url_map": "(\d*[^"]+)",.*', re.M)
def _fetch_url(url, ref=None, path=None):
opener = urllib2.build_opener()
headers = {}
if ref:
headers['Referer'] = ref
request = urllib2.Request(url, headers=headers)
handle = urllib2.urlopen(request)
if not path:
return handle.read()
sys.stdout.write('saving: ')
# write result to file
with open(path, 'wb') as out:
while True:
part = handle.read(65536)
if not part:
break
out.write(part)
sys.stdout.write('.')
sys.stdout.flush()
sys.stdout.write('\nFinished.\n')
def _extract(html):
tree = lxml.html.fromstring(html)
res = {'204': _RE_G204.findall(html)[0].replace('\\', '')}
for script in tree.findall('.//script'):
text = script.text_content()
if 'fmt_url_map' not in text:
continue
# found it, extract the urls we need
for tmp in _RE_URLS.findall(text)[0].split(','):
url_id, url = tmp.split('|')
res[url_id] = url.replace('\\', '')
break
return res
def main():
target = sys.argv[1]
dest = sys.argv[2]
html = _fetch_url(target)
res = dict(_extract(html))
# hit the 'generate_204' url first and remove it
_fetch_url(res['204'], ref=target)
del res['204']
# download the video. now i grab the first 'download' url and use it.
first = res.values()[0]
_fetch_url(first, ref=target, path=dest)
if __name__ == '__main__':
main()
运行它:
python youdown.py 'http://www.youtube.com/watch?v=Je_iqbgGXFw' stevegadd.flv
saving: ........................... finished.
答案 2 :(得分:4)
我建议使用urllib2或beautifulsoup编写自己的解析器。您可以查看DownThemAll的源代码,了解该插件如何找到视频网址