]> Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/generic.py
839530982e7db4704ece2a589420ab4ecfc31c66
[youtubedl] / youtube_dl / extractor / generic.py
1 # encoding: utf-8
2
3 from __future__ import unicode_literals
4
5 import os
6 import re
7
8 from .common import InfoExtractor
9 from .youtube import YoutubeIE
10 from ..utils import (
11 compat_urllib_error,
12 compat_urllib_parse,
13 compat_urllib_request,
14 compat_urlparse,
15
16 ExtractorError,
17 HEADRequest,
18 smuggle_url,
19 unescapeHTML,
20 unified_strdate,
21 url_basename,
22 )
23 from .brightcove import BrightcoveIE
24 from .ooyala import OoyalaIE
25
26
27 class GenericIE(InfoExtractor):
28 IE_DESC = 'Generic downloader that works on some sites'
29 _VALID_URL = r'.*'
30 IE_NAME = 'generic'
31 _TESTS = [
32 {
33 'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
34 'file': '13601338388002.mp4',
35 'md5': '6e15c93721d7ec9e9ca3fdbf07982cfd',
36 'info_dict': {
37 'uploader': 'www.hodiho.fr',
38 'title': 'R\u00e9gis plante sa Jeep',
39 }
40 },
41 # embedded vimeo video
42 {
43 'add_ie': ['Vimeo'],
44 'url': 'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
45 'file': '22444065.mp4',
46 'md5': '2903896e23df39722c33f015af0666e2',
47 'info_dict': {
48 'title': 'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
49 'uploader_id': 'skillsmatter',
50 'uploader': 'Skills Matter',
51 }
52 },
53 # bandcamp page with custom domain
54 {
55 'add_ie': ['Bandcamp'],
56 'url': 'http://bronyrock.com/track/the-pony-mash',
57 'file': '3235767654.mp3',
58 'info_dict': {
59 'title': 'The Pony Mash',
60 'uploader': 'M_Pallante',
61 },
62 'skip': 'There is a limit of 200 free downloads / month for the test song',
63 },
64 # embedded brightcove video
65 # it also tests brightcove videos that need to set the 'Referer' in the
66 # http requests
67 {
68 'add_ie': ['Brightcove'],
69 'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
70 'info_dict': {
71 'id': '2765128793001',
72 'ext': 'mp4',
73 'title': 'Le cours de bourse : l’analyse technique',
74 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
75 'uploader': 'BFM BUSINESS',
76 },
77 'params': {
78 'skip_download': True,
79 },
80 },
81 # Direct link to a video
82 {
83 'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
84 'file': 'trailer.mp4',
85 'md5': '67d406c2bcb6af27fa886f31aa934bbe',
86 'info_dict': {
87 'id': 'trailer',
88 'title': 'trailer',
89 'upload_date': '20100513',
90 }
91 },
92 # ooyala video
93 {
94 'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
95 'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
96 'info_dict': {
97 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
98 'ext': 'mp4',
99 'title': '2cc213299525360.mov', #that's what we get
100 },
101 },
102 ]
103
104 def report_download_webpage(self, video_id):
105 """Report webpage download."""
106 if not self._downloader.params.get('test', False):
107 self._downloader.report_warning('Falling back on generic information extractor.')
108 super(GenericIE, self).report_download_webpage(video_id)
109
110 def report_following_redirect(self, new_url):
111 """Report information extraction."""
112 self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
113
114 def _send_head(self, url):
115 """Check if it is a redirect, like url shorteners, in case return the new url."""
116
117 class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
118 """
119 Subclass the HTTPRedirectHandler to make it use our
120 HEADRequest also on the redirected URL
121 """
122 def redirect_request(self, req, fp, code, msg, headers, newurl):
123 if code in (301, 302, 303, 307):
124 newurl = newurl.replace(' ', '%20')
125 newheaders = dict((k,v) for k,v in req.headers.items()
126 if k.lower() not in ("content-length", "content-type"))
127 return HEADRequest(newurl,
128 headers=newheaders,
129 origin_req_host=req.get_origin_req_host(),
130 unverifiable=True)
131 else:
132 raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
133
134 class HTTPMethodFallback(compat_urllib_request.BaseHandler):
135 """
136 Fallback to GET if HEAD is not allowed (405 HTTP error)
137 """
138 def http_error_405(self, req, fp, code, msg, headers):
139 fp.read()
140 fp.close()
141
142 newheaders = dict((k,v) for k,v in req.headers.items()
143 if k.lower() not in ("content-length", "content-type"))
144 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
145 headers=newheaders,
146 origin_req_host=req.get_origin_req_host(),
147 unverifiable=True))
148
149 # Build our opener
150 opener = compat_urllib_request.OpenerDirector()
151 for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
152 HTTPMethodFallback, HEADRedirectHandler,
153 compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
154 opener.add_handler(handler())
155
156 response = opener.open(HEADRequest(url))
157 if response is None:
158 raise ExtractorError('Invalid URL protocol')
159 return response
160
161 def _real_extract(self, url):
162 parsed_url = compat_urlparse.urlparse(url)
163 if not parsed_url.scheme:
164 self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
165 return self.url_result('http://' + url)
166 video_id = os.path.splitext(url.split('/')[-1])[0]
167
168 self.to_screen('%s: Requesting header' % video_id)
169
170 try:
171 response = self._send_head(url)
172
173 # Check for redirect
174 new_url = response.geturl()
175 if url != new_url:
176 self.report_following_redirect(new_url)
177 return self.url_result(new_url)
178
179 # Check for direct link to a video
180 content_type = response.headers.get('Content-Type', '')
181 m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
182 if m:
183 upload_date = response.headers.get('Last-Modified')
184 if upload_date:
185 upload_date = unified_strdate(upload_date)
186 return {
187 'id': video_id,
188 'title': os.path.splitext(url_basename(url))[0],
189 'formats': [{
190 'format_id': m.group('format_id'),
191 'url': url,
192 'vcodec': 'none' if m.group('type') == 'audio' else None
193 }],
194 'upload_date': upload_date,
195 }
196
197 except compat_urllib_error.HTTPError:
198 # This may be a stupid server that doesn't like HEAD, our UA, or so
199 pass
200
201 try:
202 webpage = self._download_webpage(url, video_id)
203 except ValueError:
204 # since this is the last-resort InfoExtractor, if
205 # this error is thrown, it'll be thrown here
206 raise ExtractorError('Failed to download URL: %s' % url)
207
208 self.report_extraction(video_id)
209
210 # it's tempting to parse this further, but you would
211 # have to take into account all the variations like
212 # Video Title - Site Name
213 # Site Name | Video Title
214 # Video Title - Tagline | Site Name
215 # and so on and so forth; it's just not practical
216 video_title = self._html_search_regex(
217 r'(?s)<title>(.*?)</title>', webpage, 'video title',
218 default='video')
219
220 # video uploader is domain name
221 video_uploader = self._search_regex(
222 r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
223
224 # Look for BrightCove:
225 bc_url = BrightcoveIE._extract_brightcove_url(webpage)
226 if bc_url is not None:
227 self.to_screen('Brightcove video detected.')
228 surl = smuggle_url(bc_url, {'Referer': url})
229 return self.url_result(surl, 'Brightcove')
230
231 # Look for embedded (iframe) Vimeo player
232 mobj = re.search(
233 r'<iframe[^>]+?src="((?:https?:)?//player.vimeo.com/video/.+?)"', webpage)
234 if mobj:
235 player_url = unescapeHTML(mobj.group(1))
236 surl = smuggle_url(player_url, {'Referer': url})
237 return self.url_result(surl, 'Vimeo')
238
239 # Look for embedded (swf embed) Vimeo player
240 mobj = re.search(
241 r'<embed[^>]+?src="(https?://(?:www\.)?vimeo.com/moogaloop.swf.+?)"', webpage)
242 if mobj:
243 return self.url_result(mobj.group(1), 'Vimeo')
244
245 # Look for embedded YouTube player
246 matches = re.findall(r'''(?x)
247 (?:<iframe[^>]+?src=|embedSWF\(\s*)
248 (["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
249 (?:embed|v)/.+?)
250 \1''', webpage)
251 if matches:
252 urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
253 for tuppl in matches]
254 return self.playlist_result(
255 urlrs, playlist_id=video_id, playlist_title=video_title)
256
257 # Look for embedded Dailymotion player
258 matches = re.findall(
259 r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
260 if matches:
261 urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
262 for tuppl in matches]
263 return self.playlist_result(
264 urlrs, playlist_id=video_id, playlist_title=video_title)
265
266 # Look for embedded Wistia player
267 match = re.search(
268 r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
269 if match:
270 return {
271 '_type': 'url_transparent',
272 'url': unescapeHTML(match.group('url')),
273 'ie_key': 'Wistia',
274 'uploader': video_uploader,
275 'title': video_title,
276 'id': video_id,
277 }
278
279 # Look for embedded blip.tv player
280 mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
281 if mobj:
282 return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
283 mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
284 if mobj:
285 return self.url_result(mobj.group(1), 'BlipTV')
286
287 # Look for Bandcamp pages with custom domain
288 mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
289 if mobj is not None:
290 burl = unescapeHTML(mobj.group(1))
291 # Don't set the extractor because it can be a track url or an album
292 return self.url_result(burl)
293
294 # Look for embedded Vevo player
295 mobj = re.search(
296 r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
297 if mobj is not None:
298 return self.url_result(mobj.group('url'))
299
300 # Look for Ooyala videos
301 mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage)
302 if mobj is not None:
303 return OoyalaIE._build_url_result(mobj.group(1))
304
305 # Look for Aparat videos
306 mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
307 if mobj is not None:
308 return self.url_result(mobj.group(1), 'Aparat')
309
310 # Look for MPORA videos
311 mobj = re.search(r'<iframe .*?src="(http://mpora\.com/videos/[^"]+)"', webpage)
312 if mobj is not None:
313 return self.url_result(mobj.group(1), 'Mpora')
314
315 # Look for embedded Novamov player
316 mobj = re.search(
317 r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?novamov\.com/embed\.php.+?)\1', webpage)
318 if mobj is not None:
319 return self.url_result(mobj.group('url'), 'Novamov')
320
321 # Start with something easy: JW Player in SWFObject
322 mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
323 if mobj is None:
324 # Look for gorilla-vid style embedding
325 mobj = re.search(r'(?s)jw_plugins.*?file:\s*["\'](.*?)["\']', webpage)
326 if mobj is None:
327 # Broaden the search a little bit
328 mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
329 if mobj is None:
330 # Broaden the search a little bit: JWPlayer JS loader
331 mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
332 if mobj is None:
333 # Try to find twitter cards info
334 mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
335 if mobj is None:
336 # We look for Open Graph info:
337 # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
338 m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
339 # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
340 if m_video_type is not None:
341 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
342 if mobj is None:
343 # HTML5 video
344 mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
345 if mobj is None:
346 raise ExtractorError('Unsupported URL: %s' % url)
347
348 # It's possible that one of the regexes
349 # matched, but returned an empty group:
350 if mobj.group(1) is None:
351 raise ExtractorError('Did not find a valid video URL at %s' % url)
352
353 video_url = mobj.group(1)
354 video_url = compat_urlparse.urljoin(url, video_url)
355 video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
356
357 # Sometimes, jwplayer extraction will result in a YouTube URL
358 if YoutubeIE.suitable(video_url):
359 return self.url_result(video_url, 'Youtube')
360
361 # here's a fun little line of code for you:
362 video_id = os.path.splitext(video_id)[0]
363
364 return {
365 'id': video_id,
366 'url': video_url,
367 'uploader': video_uploader,
368 'title': video_title,
369 }