yt-dlp/youtube_dl/extractor/generic.py

406 lines
17 KiB
Python
Raw Normal View History

# encoding: utf-8
from __future__ import unicode_literals
2013-06-23 18:31:45 +00:00
import os
import re
from .common import InfoExtractor
from .youtube import YoutubeIE
2013-06-23 18:31:45 +00:00
from ..utils import (
compat_urllib_error,
compat_urllib_parse,
compat_urllib_request,
2013-08-28 10:47:27 +00:00
compat_urlparse,
2013-06-23 18:31:45 +00:00
ExtractorError,
2013-12-20 16:05:28 +00:00
HEADRequest,
smuggle_url,
unescapeHTML,
unified_strdate,
url_basename,
2013-06-23 18:31:45 +00:00
)
from .brightcove import BrightcoveIE
from .ooyala import OoyalaIE
2013-06-23 18:31:45 +00:00
2013-06-23 18:31:45 +00:00
class GenericIE(InfoExtractor):
IE_DESC = 'Generic downloader that works on some sites'
2013-06-23 18:31:45 +00:00
_VALID_URL = r'.*'
IE_NAME = 'generic'
_TESTS = [
{
'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
'file': '13601338388002.mp4',
'md5': '6e15c93721d7ec9e9ca3fdbf07982cfd',
'info_dict': {
'uploader': 'www.hodiho.fr',
'title': 'R\u00e9gis plante sa Jeep',
}
},
# embedded vimeo video
{
'add_ie': ['Vimeo'],
'url': 'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
'file': '22444065.mp4',
'md5': '2903896e23df39722c33f015af0666e2',
'info_dict': {
'title': 'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
'uploader_id': 'skillsmatter',
'uploader': 'Skills Matter',
}
},
# bandcamp page with custom domain
{
'add_ie': ['Bandcamp'],
'url': 'http://bronyrock.com/track/the-pony-mash',
'file': '3235767654.mp3',
'info_dict': {
'title': 'The Pony Mash',
'uploader': 'M_Pallante',
},
'skip': 'There is a limit of 200 free downloads / month for the test song',
},
# embedded brightcove video
# it also tests brightcove videos that need to set the 'Referer' in the
# http requests
{
'add_ie': ['Brightcove'],
'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
'info_dict': {
'id': '2765128793001',
'ext': 'mp4',
'title': 'Le cours de bourse : lanalyse technique',
'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
'uploader': 'BFM BUSINESS',
},
'params': {
'skip_download': True,
},
},
2014-01-28 02:35:32 +00:00
{
# https://github.com/rg3/youtube-dl/issues/2253
'url': 'http://bcove.me/i6nfkrc3',
'file': '3101154703001.mp4',
'md5': '0ba9446db037002366bab3b3eb30c88c',
'info_dict': {
'title': 'Still no power',
'uploader': 'thestar.com',
'description': 'Mississauga resident David Farmer is still out of power as a result of the ice storm a month ago. To keep the house warm, Farmer cuts wood from his property for a wood burning stove downstairs.',
},
'add_ie': ['Brightcove'],
},
# Direct link to a video
{
'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
'file': 'trailer.mp4',
'md5': '67d406c2bcb6af27fa886f31aa934bbe',
'info_dict': {
'id': 'trailer',
'title': 'trailer',
'upload_date': '20100513',
}
},
# ooyala video
{
'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
2014-01-21 00:40:34 +00:00
'file': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ.mp4',
'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
'info_dict': {
'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
'ext': 'mp4',
2014-01-21 00:40:34 +00:00
'title': '2cc213299525360.mov', # that's what we get
},
},
]
2013-06-23 18:31:45 +00:00
def report_download_webpage(self, video_id):
"""Report webpage download."""
if not self._downloader.params.get('test', False):
self._downloader.report_warning('Falling back on generic information extractor.')
2013-06-23 18:31:45 +00:00
super(GenericIE, self).report_download_webpage(video_id)
def report_following_redirect(self, new_url):
"""Report information extraction."""
self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
2013-06-23 18:31:45 +00:00
def _send_head(self, url):
2013-06-23 18:31:45 +00:00
"""Check if it is a redirect, like url shorteners, in case return the new url."""
class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
"""
Subclass the HTTPRedirectHandler to make it use our
2013-12-20 16:05:28 +00:00
HEADRequest also on the redirected URL
2013-06-23 18:31:45 +00:00
"""
def redirect_request(self, req, fp, code, msg, headers, newurl):
if code in (301, 302, 303, 307):
newurl = newurl.replace(' ', '%20')
newheaders = dict((k,v) for k,v in req.headers.items()
if k.lower() not in ("content-length", "content-type"))
2013-12-20 16:05:28 +00:00
return HEADRequest(newurl,
2013-06-23 18:31:45 +00:00
headers=newheaders,
origin_req_host=req.get_origin_req_host(),
unverifiable=True)
else:
raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
class HTTPMethodFallback(compat_urllib_request.BaseHandler):
"""
Fallback to GET if HEAD is not allowed (405 HTTP error)
"""
def http_error_405(self, req, fp, code, msg, headers):
fp.read()
fp.close()
newheaders = dict((k,v) for k,v in req.headers.items()
if k.lower() not in ("content-length", "content-type"))
return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
headers=newheaders,
origin_req_host=req.get_origin_req_host(),
unverifiable=True))
# Build our opener
opener = compat_urllib_request.OpenerDirector()
for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
HTTPMethodFallback, HEADRedirectHandler,
compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
opener.add_handler(handler())
2013-12-20 16:05:28 +00:00
response = opener.open(HEADRequest(url))
2013-06-23 18:31:45 +00:00
if response is None:
raise ExtractorError('Invalid URL protocol')
return response
2013-06-23 18:31:45 +00:00
def _real_extract(self, url):
parsed_url = compat_urlparse.urlparse(url)
if not parsed_url.scheme:
default_search = self._downloader.params.get('default_search')
if default_search is None:
default_search = 'auto'
if default_search == 'auto':
if '/' in url:
self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
return self.url_result('http://' + url)
else:
return self.url_result('ytsearch:' + url)
else:
assert ':' in default_search
return self.url_result(default_search + url)
video_id = os.path.splitext(url.split('/')[-1])[0]
self.to_screen('%s: Requesting header' % video_id)
try:
response = self._send_head(url)
# Check for redirect
new_url = response.geturl()
if url != new_url:
self.report_following_redirect(new_url)
return self.url_result(new_url)
# Check for direct link to a video
content_type = response.headers.get('Content-Type', '')
m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
if m:
upload_date = response.headers.get('Last-Modified')
if upload_date:
upload_date = unified_strdate(upload_date)
return {
'id': video_id,
'title': os.path.splitext(url_basename(url))[0],
'formats': [{
'format_id': m.group('format_id'),
'url': url,
'vcodec': 'none' if m.group('type') == 'audio' else None
}],
'upload_date': upload_date,
}
except compat_urllib_error.HTTPError:
# This may be a stupid server that doesn't like HEAD, our UA, or so
pass
2013-06-23 18:31:45 +00:00
try:
webpage = self._download_webpage(url, video_id)
except ValueError:
# since this is the last-resort InfoExtractor, if
# this error is thrown, it'll be thrown here
raise ExtractorError('Failed to download URL: %s' % url)
2013-06-23 18:31:45 +00:00
self.report_extraction(video_id)
# it's tempting to parse this further, but you would
# have to take into account all the variations like
# Video Title - Site Name
# Site Name | Video Title
# Video Title - Tagline | Site Name
# and so on and so forth; it's just not practical
2013-12-06 08:15:04 +00:00
video_title = self._html_search_regex(
r'(?s)<title>(.*?)</title>', webpage, 'video title',
default='video')
2013-12-06 08:15:04 +00:00
# video uploader is domain name
video_uploader = self._search_regex(
r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
2013-08-26 19:29:31 +00:00
# Look for BrightCove:
bc_url = BrightcoveIE._extract_brightcove_url(webpage)
if bc_url is not None:
self.to_screen('Brightcove video detected.')
surl = smuggle_url(bc_url, {'Referer': url})
return self.url_result(surl, 'Brightcove')
# Look for embedded (iframe) Vimeo player
mobj = re.search(
2014-01-08 21:42:52 +00:00
r'<iframe[^>]+?src="((?:https?:)?//player.vimeo.com/video/.+?)"', webpage)
if mobj:
player_url = unescapeHTML(mobj.group(1))
surl = smuggle_url(player_url, {'Referer': url})
return self.url_result(surl, 'Vimeo')
# Look for embedded (swf embed) Vimeo player
mobj = re.search(
r'<embed[^>]+?src="(https?://(?:www\.)?vimeo.com/moogaloop.swf.+?)"', webpage)
if mobj:
return self.url_result(mobj.group(1), 'Vimeo')
# Look for embedded YouTube player
matches = re.findall(r'''(?x)
(?:<iframe[^>]+?src=|embedSWF\(\s*)
(["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
(?:embed|v)/.+?)
\1''', webpage)
if matches:
urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
for tuppl in matches]
return self.playlist_result(
urlrs, playlist_id=video_id, playlist_title=video_title)
# Look for embedded Dailymotion player
matches = re.findall(
2013-12-06 08:15:04 +00:00
r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
if matches:
urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
for tuppl in matches]
return self.playlist_result(
urlrs, playlist_id=video_id, playlist_title=video_title)
2013-12-06 08:15:04 +00:00
# Look for embedded Wistia player
match = re.search(
r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
if match:
return {
'_type': 'url_transparent',
'url': unescapeHTML(match.group('url')),
'ie_key': 'Wistia',
'uploader': video_uploader,
'title': video_title,
'id': video_id,
}
# Look for embedded blip.tv player
mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
if mobj:
return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
if mobj:
return self.url_result(mobj.group(1), 'BlipTV')
# Look for Bandcamp pages with custom domain
mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
if mobj is not None:
burl = unescapeHTML(mobj.group(1))
# Don't set the extractor because it can be a track url or an album
return self.url_result(burl)
# Look for embedded Vevo player
mobj = re.search(
r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
if mobj is not None:
return self.url_result(mobj.group('url'))
# Look for Ooyala videos
mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage)
if mobj is not None:
return OoyalaIE._build_url_result(mobj.group(1))
2013-12-20 16:05:28 +00:00
# Look for Aparat videos
2014-01-07 10:04:27 +00:00
mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
2013-12-20 16:05:28 +00:00
if mobj is not None:
return self.url_result(mobj.group(1), 'Aparat')
2014-01-07 07:07:46 +00:00
# Look for MPORA videos
2014-01-07 10:04:27 +00:00
mobj = re.search(r'<iframe .*?src="(http://mpora\.com/videos/[^"]+)"', webpage)
2014-01-07 07:07:46 +00:00
if mobj is not None:
return self.url_result(mobj.group(1), 'Mpora')
2014-01-08 01:11:46 +00:00
2014-01-08 01:07:11 +00:00
# Look for embedded Novamov player
mobj = re.search(
r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?novamov\.com/embed\.php.+?)\1', webpage)
if mobj is not None:
return self.url_result(mobj.group('url'), 'Novamov')
2014-01-07 07:07:46 +00:00
# Look for embedded Facebook player
mobj = re.search(
2014-01-27 04:47:30 +00:00
r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
if mobj is not None:
return self.url_result(mobj.group('url'), 'Facebook')
2014-01-27 04:47:30 +00:00
# Look for embedded Huffington Post player
mobj = re.search(
r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live.huffingtonpost\.com/.+?)\1', webpage)
if mobj is not None:
return self.url_result(mobj.group('url'), 'HuffPost')
2013-06-23 18:31:45 +00:00
# Start with something easy: JW Player in SWFObject
mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
if mobj is None:
# Look for gorilla-vid style embedding
mobj = re.search(r'(?s)jw_plugins.*?file:\s*["\'](.*?)["\']', webpage)
2013-06-23 18:31:45 +00:00
if mobj is None:
# Broaden the search a little bit
mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
if mobj is None:
# Broaden the search a little bit: JWPlayer JS loader
mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
2013-06-23 18:31:45 +00:00
if mobj is None:
# Try to find twitter cards info
mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
if mobj is None:
# We look for Open Graph info:
# We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
# We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
if m_video_type is not None:
mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
2013-08-21 02:32:22 +00:00
if mobj is None:
# HTML5 video
mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
2013-06-23 18:31:45 +00:00
if mobj is None:
raise ExtractorError('Unsupported URL: %s' % url)
2013-06-23 18:31:45 +00:00
# It's possible that one of the regexes
# matched, but returned an empty group:
if mobj.group(1) is None:
raise ExtractorError('Did not find a valid video URL at %s' % url)
2013-06-23 18:31:45 +00:00
video_url = mobj.group(1)
2013-08-28 10:47:27 +00:00
video_url = compat_urlparse.urljoin(url, video_url)
video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
2013-06-23 18:31:45 +00:00
# Sometimes, jwplayer extraction will result in a YouTube URL
if YoutubeIE.suitable(video_url):
return self.url_result(video_url, 'Youtube')
2013-06-23 18:31:45 +00:00
# here's a fun little line of code for you:
video_id = os.path.splitext(video_id)[0]
return {
'id': video_id,
'url': video_url,
2013-06-23 18:31:45 +00:00
'uploader': video_uploader,
'title': video_title,
}