diff options
Diffstat (limited to 'youtube_dl/extractor/common.py')
| -rw-r--r-- | youtube_dl/extractor/common.py | 286 |
1 files changed, 224 insertions, 62 deletions
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index b9014fc23..16ae4b98f 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -14,10 +14,13 @@ import xml.etree.ElementTree from ..compat import ( compat_cookiejar, + compat_cookies, compat_HTTPError, compat_http_client, compat_urllib_error, + compat_urllib_parse, compat_urllib_parse_urlparse, + compat_urllib_request, compat_urlparse, compat_str, ) @@ -35,6 +38,9 @@ from ..utils import ( RegexNotFoundError, sanitize_filename, unescapeHTML, + url_basename, + xpath_text, + xpath_with_ns, ) @@ -65,7 +71,7 @@ class InfoExtractor(object): Potential fields: * url Mandatory. The URL of the video file - * ext Will be calculated from url if missing + * ext Will be calculated from URL if missing * format A human-readable description of the format ("mp4 container with h264/opus"). Calculated from the format_id, width, height. @@ -155,7 +161,7 @@ class InfoExtractor(object): lower to higher preference, each element is a dictionary with the "ext" entry and one of: * "data": The subtitles file contents - * "url": A url pointing to the subtitles file + * "url": A URL pointing to the subtitles file automatic_captions: Like 'subtitles', used by the YoutubeIE for automatically generated captions duration: Length of the video in seconds, as an integer. @@ -176,13 +182,18 @@ class InfoExtractor(object): Set to "root" to indicate that this is a comment to the original video. age_limit: Age restriction for the video, as an integer (years) - webpage_url: The url to the video webpage, if given to youtube-dl it + webpage_url: The URL to the video webpage, if given to youtube-dl it should allow to get the same result again. (It will be set by YoutubeDL if it's missing) categories: A list of categories that the video falls in, for example ["Sports", "Berlin"] + tags: A list of tags assigned to the video, e.g. ["sweden", "pop music"] is_live: True, False, or None (=unknown). Whether this video is a live stream that goes on instead of a fixed-length video. + start_time: Time in seconds where the reproduction should start, as + specified in the URL. + end_time: Time in seconds where the reproduction should end, as + specified in the URL. Unless mentioned otherwise, the fields should be Unicode strings. @@ -193,8 +204,8 @@ class InfoExtractor(object): There must be a key "entries", which is a list, an iterable, or a PagedList object, each element of which is a valid dictionary by this specification. - Additionally, playlists can have "title" and "id" attributes with the same - semantics as videos (see above). + Additionally, playlists can have "title", "description" and "id" attributes + with the same semantics as videos (see above). _type "multi_video" indicates that there are multiple videos that @@ -501,7 +512,7 @@ class InfoExtractor(object): # Methods for following #608 @staticmethod def url_result(url, ie=None, video_id=None, video_title=None): - """Returns a url that points to a page that should be processed""" + """Returns a URL that points to a page that should be processed""" # TODO: ie should be the class used for getting the info video_info = {'_type': 'url', 'url': url, @@ -626,6 +637,12 @@ class InfoExtractor(object): template % (content_re, property_re), ] + @staticmethod + def _meta_regex(prop): + return r'''(?isx)<meta + (?=[^>]+(?:itemprop|name|property|id|http-equiv)=(["\']?)%s\1) + [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop) + def _og_search_property(self, prop, html, name=None, **kargs): if name is None: name = 'OpenGraph %s' % prop @@ -635,7 +652,7 @@ class InfoExtractor(object): return unescapeHTML(escaped) def _og_search_thumbnail(self, html, **kargs): - return self._og_search_property('image', html, 'thumbnail url', fatal=False, **kargs) + return self._og_search_property('image', html, 'thumbnail URL', fatal=False, **kargs) def _og_search_description(self, html, **kargs): return self._og_search_property('description', html, fatal=False, **kargs) @@ -656,9 +673,7 @@ class InfoExtractor(object): if display_name is None: display_name = name return self._html_search_regex( - r'''(?isx)<meta - (?=[^>]+(?:itemprop|name|property)=(["\']?)%s\1) - [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(name), + self._meta_regex(name), html, display_name, fatal=fatal, group='content', **kwargs) def _dc_search_uploader(self, html): @@ -967,69 +982,210 @@ class InfoExtractor(object): self._sort_formats(formats) return formats - # TODO: improve extraction - def _extract_smil_formats(self, smil_url, video_id, fatal=True): - smil = self._download_xml( - smil_url, video_id, 'Downloading SMIL file', - 'Unable to download SMIL file', fatal=fatal) + @staticmethod + def _xpath_ns(path, namespace=None): + if not namespace: + return path + out = [] + for c in path.split('/'): + if not c or c == '.': + out.append(c) + else: + out.append('{%s}%s' % (namespace, c)) + return '/'.join(out) + + def _extract_smil_formats(self, smil_url, video_id, fatal=True, f4m_params=None): + smil = self._download_smil(smil_url, video_id, fatal=fatal) + if smil is False: assert not fatal return [] - base = smil.find('./head/meta').get('base') + namespace = self._parse_smil_namespace(smil) + + return self._parse_smil_formats( + smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params) + + def _extract_smil_info(self, smil_url, video_id, fatal=True, f4m_params=None): + smil = self._download_smil(smil_url, video_id, fatal=fatal) + if smil is False: + return {} + return self._parse_smil(smil, smil_url, video_id, f4m_params=f4m_params) + + def _download_smil(self, smil_url, video_id, fatal=True): + return self._download_xml( + smil_url, video_id, 'Downloading SMIL file', + 'Unable to download SMIL file', fatal=fatal) + + def _parse_smil(self, smil, smil_url, video_id, f4m_params=None): + namespace = self._parse_smil_namespace(smil) + + formats = self._parse_smil_formats( + smil, smil_url, video_id, namespace=namespace, f4m_params=f4m_params) + subtitles = self._parse_smil_subtitles(smil, namespace=namespace) + + video_id = os.path.splitext(url_basename(smil_url))[0] + title = None + description = None + for meta in smil.findall(self._xpath_ns('./head/meta', namespace)): + name = meta.attrib.get('name') + content = meta.attrib.get('content') + if not name or not content: + continue + if not title and name == 'title': + title = content + elif not description and name in ('description', 'abstract'): + description = content + + return { + 'id': video_id, + 'title': title or video_id, + 'description': description, + 'formats': formats, + 'subtitles': subtitles, + } + + def _parse_smil_namespace(self, smil): + return self._search_regex( + r'(?i)^{([^}]+)?}smil$', smil.tag, 'namespace', default=None) + + def _parse_smil_formats(self, smil, smil_url, video_id, namespace=None, f4m_params=None): + base = smil_url + for meta in smil.findall(self._xpath_ns('./head/meta', namespace)): + b = meta.get('base') or meta.get('httpBase') + if b: + base = b + break formats = [] rtmp_count = 0 - if smil.findall('./body/seq/video'): - video = smil.findall('./body/seq/video')[0] - fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count) - formats.extend(fmts) - else: - for video in smil.findall('./body/switch/video'): - fmts, rtmp_count = self._parse_smil_video(video, video_id, base, rtmp_count) - formats.extend(fmts) + http_count = 0 + + videos = smil.findall(self._xpath_ns('.//video', namespace)) + for video in videos: + src = video.get('src') + if not src: + continue + + bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000) + filesize = int_or_none(video.get('size') or video.get('fileSize')) + width = int_or_none(video.get('width')) + height = int_or_none(video.get('height')) + proto = video.get('proto') + ext = video.get('ext') + src_ext = determine_ext(src) + streamer = video.get('streamer') or base + + if proto == 'rtmp' or streamer.startswith('rtmp'): + rtmp_count += 1 + formats.append({ + 'url': streamer, + 'play_path': src, + 'ext': 'flv', + 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate), + 'tbr': bitrate, + 'filesize': filesize, + 'width': width, + 'height': height, + }) + continue + + src_url = src if src.startswith('http') else compat_urlparse.urljoin(base, src) + + if proto == 'm3u8' or src_ext == 'm3u8': + formats.extend(self._extract_m3u8_formats( + src_url, video_id, ext or 'mp4', m3u8_id='hls')) + continue + + if src_ext == 'f4m': + f4m_url = src_url + if not f4m_params: + f4m_params = { + 'hdcore': '3.2.0', + 'plugin': 'flowplayer-3.2.0.1', + } + f4m_url += '&' if '?' in f4m_url else '?' + f4m_url += compat_urllib_parse.urlencode(f4m_params) + formats.extend(self._extract_f4m_formats(f4m_url, video_id, f4m_id='hds')) + continue + + if src_url.startswith('http'): + http_count += 1 + formats.append({ + 'url': src_url, + 'ext': ext or src_ext or 'flv', + 'format_id': 'http-%d' % (bitrate or http_count), + 'tbr': bitrate, + 'filesize': filesize, + 'width': width, + 'height': height, + }) + continue self._sort_formats(formats) return formats - def _parse_smil_video(self, video, video_id, base, rtmp_count): - src = video.get('src') - if not src: - return [], rtmp_count - bitrate = int_or_none(video.get('system-bitrate') or video.get('systemBitrate'), 1000) - width = int_or_none(video.get('width')) - height = int_or_none(video.get('height')) - proto = video.get('proto') - if not proto: - if base: - if base.startswith('rtmp'): - proto = 'rtmp' - elif base.startswith('http'): - proto = 'http' - ext = video.get('ext') - if proto == 'm3u8': - return self._extract_m3u8_formats(src, video_id, ext), rtmp_count - elif proto == 'rtmp': - rtmp_count += 1 - streamer = video.get('streamer') or base - return ([{ - 'url': streamer, - 'play_path': src, - 'ext': 'flv', - 'format_id': 'rtmp-%d' % (rtmp_count if bitrate is None else bitrate), - 'tbr': bitrate, - 'width': width, - 'height': height, - }], rtmp_count) - elif proto.startswith('http'): - return ([{ - 'url': base + src, - 'ext': ext or 'flv', - 'tbr': bitrate, - 'width': width, - 'height': height, - }], rtmp_count) + def _parse_smil_subtitles(self, smil, namespace=None): + subtitles = {} + for num, textstream in enumerate(smil.findall(self._xpath_ns('.//textstream', namespace))): + src = textstream.get('src') + if not src: + continue + ext = textstream.get('ext') or determine_ext(src) + if not ext: + type_ = textstream.get('type') + if type_ == 'text/srt': + ext = 'srt' + lang = textstream.get('systemLanguage') or textstream.get('systemLanguageName') + subtitles.setdefault(lang, []).append({ + 'url': src, + 'ext': ext, + }) + return subtitles + + def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True): + xspf = self._download_xml( + playlist_url, playlist_id, 'Downloading xpsf playlist', + 'Unable to download xspf manifest', fatal=fatal) + if xspf is False: + return [] + return self._parse_xspf(xspf, playlist_id) + + def _parse_xspf(self, playlist, playlist_id): + NS_MAP = { + 'xspf': 'http://xspf.org/ns/0/', + 's1': 'http://static.streamone.nl/player/ns/0', + } + + entries = [] + for track in playlist.findall(xpath_with_ns('./xspf:trackList/xspf:track', NS_MAP)): + title = xpath_text( + track, xpath_with_ns('./xspf:title', NS_MAP), 'title', default=playlist_id) + description = xpath_text( + track, xpath_with_ns('./xspf:annotation', NS_MAP), 'description') + thumbnail = xpath_text( + track, xpath_with_ns('./xspf:image', NS_MAP), 'thumbnail') + duration = float_or_none( + xpath_text(track, xpath_with_ns('./xspf:duration', NS_MAP), 'duration'), 1000) + + formats = [{ + 'url': location.text, + 'format_id': location.get(xpath_with_ns('s1:label', NS_MAP)), + 'width': int_or_none(location.get(xpath_with_ns('s1:width', NS_MAP))), + 'height': int_or_none(location.get(xpath_with_ns('s1:height', NS_MAP))), + } for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP))] + self._sort_formats(formats) + + entries.append({ + 'id': playlist_id, + 'title': title, + 'description': description, + 'thumbnail': thumbnail, + 'duration': duration, + 'formats': formats, + }) + return entries def _live_title(self, name): """ Generate the title for a live video """ @@ -1065,6 +1221,12 @@ class InfoExtractor(object): None, '/', True, False, expire_time, '', None, None, None) self._downloader.cookiejar.set_cookie(cookie) + def _get_cookies(self, url): + """ Return a compat_cookies.SimpleCookie with the cookies for the url """ + req = compat_urllib_request.Request(url) + self._downloader.cookiejar.add_cookie_header(req) + return compat_cookies.SimpleCookie(req.get_header('Cookie')) + def get_testcases(self, include_onlymatching=False): t = getattr(self, '_TEST', None) if t: @@ -1116,7 +1278,7 @@ class InfoExtractor(object): class SearchInfoExtractor(InfoExtractor): """ Base class for paged search queries extractors. - They accept urls in the format _SEARCH_KEY(|all|[0-9]):{query} + They accept URLs in the format _SEARCH_KEY(|all|[0-9]):{query} Instances should define _SEARCH_KEY and _MAX_RESULTS. """ |