X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/99a6f8745584b59fe0fa1f5e81f6c3667bd3fb2b..c39143087e10ba3fe3c3a309b847354b82d6125e:/youtube_dl/extractor/common.py diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index deafb48..9e7febc 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -19,6 +19,7 @@ from ..compat import ( compat_cookies, compat_etree_fromstring, compat_getpass, + compat_integer_types, compat_http_client, compat_os_name, compat_str, @@ -51,6 +52,7 @@ from ..utils import ( GeoUtils, int_or_none, js_to_json, + JSON_LD_RE, mimetype2ext, orderedSet, parse_codecs, @@ -67,6 +69,7 @@ from ..utils import ( update_url_query, urljoin, url_basename, + url_or_none, xpath_element, xpath_text, xpath_with_ns, @@ -174,6 +177,8 @@ class InfoExtractor(object): width : height ratio as float. * no_resume The server does not support resuming the (HTTP or RTMP) download. Boolean. + * downloader_options A dictionary of downloader options as + described in FileDownloader url: Final video URL. ext: Video filename extension. @@ -207,6 +212,11 @@ class InfoExtractor(object): If not explicitly set, calculated from timestamp. uploader_id: Nickname or id of the video uploader. uploader_url: Full URL to a personal webpage of the video uploader. + channel: Full name of the channel the video is uploaded on. + Note that channel fields may or may not repeat uploader + fields. This depends on a particular extractor. + channel_id: Id of the channel. + channel_url: Full URL to a channel webpage. location: Physical location where the video was filmed. subtitles: The available subtitles as a dictionary in the format {tag: subformats}. "tag" is usually a language code, and @@ -337,15 +347,17 @@ class InfoExtractor(object): _GEO_BYPASS attribute may be set to False in order to disable geo restriction bypass mechanisms for a particular extractor. Though it won't disable explicit geo restriction bypass based on - country code provided with geo_bypass_country. (experimental) + country code provided with geo_bypass_country. _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted countries for this extractor. One of these countries will be used by geo restriction bypass mechanism right away in order to bypass - geo restriction, of course, if the mechanism is not disabled. (experimental) + geo restriction, of course, if the mechanism is not disabled. - NB: both these geo attributes are experimental and may change in future - or be completely removed. + _GEO_IP_BLOCKS attribute may contain a list of presumably geo unrestricted + IP blocks in CIDR notation for this extractor. One of these IP blocks + will be used by geo restriction bypass mechanism similarly + to _GEO_COUNTRIES. Finally, the _WORKING attribute should be set to False for broken IEs in order to warn the users and skip the tests. @@ -356,6 +368,7 @@ class InfoExtractor(object): _x_forwarded_for_ip = None _GEO_BYPASS = True _GEO_COUNTRIES = None + _GEO_IP_BLOCKS = None _WORKING = True def __init__(self, downloader=None): @@ -390,12 +403,15 @@ class InfoExtractor(object): def initialize(self): """Initializes an instance (authentication, etc).""" - self._initialize_geo_bypass(self._GEO_COUNTRIES) + self._initialize_geo_bypass({ + 'countries': self._GEO_COUNTRIES, + 'ip_blocks': self._GEO_IP_BLOCKS, + }) if not self._ready: self._real_initialize() self._ready = True - def _initialize_geo_bypass(self, countries): + def _initialize_geo_bypass(self, geo_bypass_context): """ Initialize geo restriction bypass mechanism. @@ -406,28 +422,82 @@ class InfoExtractor(object): HTTP requests. This method will be used for initial geo bypass mechanism initialization - during the instance initialization with _GEO_COUNTRIES. + during the instance initialization with _GEO_COUNTRIES and + _GEO_IP_BLOCKS. - You may also manually call it from extractor's code if geo countries + You may also manually call it from extractor's code if geo bypass information is not available beforehand (e.g. obtained during - extraction) or due to some another reason. + extraction) or due to some other reason. In this case you should pass + this information in geo bypass context passed as first argument. It may + contain following fields: + + countries: List of geo unrestricted countries (similar + to _GEO_COUNTRIES) + ip_blocks: List of geo unrestricted IP blocks in CIDR notation + (similar to _GEO_IP_BLOCKS) + """ if not self._x_forwarded_for_ip: - country_code = self._downloader.params.get('geo_bypass_country', None) - # If there is no explicit country for geo bypass specified and - # the extractor is known to be geo restricted let's fake IP - # as X-Forwarded-For right away. - if (not country_code and - self._GEO_BYPASS and - self._downloader.params.get('geo_bypass', True) and - countries): - country_code = random.choice(countries) - if country_code: - self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code) + + # Geo bypass mechanism is explicitly disabled by user + if not self._downloader.params.get('geo_bypass', True): + return + + if not geo_bypass_context: + geo_bypass_context = {} + + # Backward compatibility: previously _initialize_geo_bypass + # expected a list of countries, some 3rd party code may still use + # it this way + if isinstance(geo_bypass_context, (list, tuple)): + geo_bypass_context = { + 'countries': geo_bypass_context, + } + + # The whole point of geo bypass mechanism is to fake IP + # as X-Forwarded-For HTTP header based on some IP block or + # country code. + + # Path 1: bypassing based on IP block in CIDR notation + + # Explicit IP block specified by user, use it right away + # regardless of whether extractor is geo bypassable or not + ip_block = self._downloader.params.get('geo_bypass_ip_block', None) + + # Otherwise use random IP block from geo bypass context but only + # if extractor is known as geo bypassable + if not ip_block: + ip_blocks = geo_bypass_context.get('ip_blocks') + if self._GEO_BYPASS and ip_blocks: + ip_block = random.choice(ip_blocks) + + if ip_block: + self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block) + if self._downloader.params.get('verbose', False): + self._downloader.to_screen( + '[debug] Using fake IP %s as X-Forwarded-For.' + % self._x_forwarded_for_ip) + return + + # Path 2: bypassing based on country code + + # Explicit country code specified by user, use it right away + # regardless of whether extractor is geo bypassable or not + country = self._downloader.params.get('geo_bypass_country', None) + + # Otherwise use random country code from geo bypass context but + # only if extractor is known as geo bypassable + if not country: + countries = geo_bypass_context.get('countries') + if self._GEO_BYPASS and countries: + country = random.choice(countries) + + if country: + self._x_forwarded_for_ip = GeoUtils.random_ipv4(country) if self._downloader.params.get('verbose', False): self._downloader.to_screen( '[debug] Using fake IP %s (%s) as X-Forwarded-For.' - % (self._x_forwarded_for_ip, country_code.upper())) + % (self._x_forwarded_for_ip, country.upper())) def extract(self, url): """Extracts URL information and returns it in list of dicts.""" @@ -486,8 +556,26 @@ class InfoExtractor(object): def IE_NAME(self): return compat_str(type(self).__name__[:-2]) - def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}): - """ Returns the response handle """ + @staticmethod + def __can_accept_status_code(err, expected_status): + assert isinstance(err, compat_urllib_error.HTTPError) + if expected_status is None: + return False + if isinstance(expected_status, compat_integer_types): + return err.code == expected_status + elif isinstance(expected_status, (list, tuple)): + return err.code in expected_status + elif callable(expected_status): + return expected_status(err.code) is True + else: + assert False + + def _request_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, data=None, headers={}, query={}, expected_status=None): + """ + Return the response handle. + + See _download_webpage docstring for arguments specification. + """ if note is None: self.report_download_webpage(video_id) elif note is not False: @@ -516,6 +604,15 @@ class InfoExtractor(object): try: return self._downloader.urlopen(url_or_request) except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: + if isinstance(err, compat_urllib_error.HTTPError): + if self.__can_accept_status_code(err, expected_status): + # Retain reference to error to prevent file object from + # being closed before it can be read. Works around the + # effects of + # introduced in Python 3.4.1. + err.fp._error = err + return err.fp + if errnote is False: return False if errnote is None: @@ -528,13 +625,17 @@ class InfoExtractor(object): self._downloader.report_warning(errmsg) return False - def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}): - """ Returns a tuple (page content as string, URL handle) """ + def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): + """ + Return a tuple (page content as string, URL handle). + + See _download_webpage docstring for arguments specification. + """ # Strip hashes from the URL (#1038) if isinstance(url_or_request, (compat_str, str)): url_or_request = url_or_request.partition('#')[0] - urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query) + urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query, expected_status=expected_status) if urlh is False: assert not fatal return False @@ -623,13 +724,52 @@ class InfoExtractor(object): return content - def _download_webpage(self, url_or_request, video_id, note=None, errnote=None, fatal=True, tries=1, timeout=5, encoding=None, data=None, headers={}, query={}): - """ Returns the data of the page as a string """ + def _download_webpage( + self, url_or_request, video_id, note=None, errnote=None, + fatal=True, tries=1, timeout=5, encoding=None, data=None, + headers={}, query={}, expected_status=None): + """ + Return the data of the page as a string. + + Arguments: + url_or_request -- plain text URL as a string or + a compat_urllib_request.Requestobject + video_id -- Video/playlist/item identifier (string) + + Keyword arguments: + note -- note printed before downloading (string) + errnote -- note printed in case of an error (string) + fatal -- flag denoting whether error should be considered fatal, + i.e. whether it should cause ExtractionError to be raised, + otherwise a warning will be reported and extraction continued + tries -- number of tries + timeout -- sleep interval between tries + encoding -- encoding for a page content decoding, guessed automatically + when not explicitly specified + data -- POST data (bytes) + headers -- HTTP headers (dict) + query -- URL query (dict) + expected_status -- allows to accept failed HTTP requests (non 2xx + status code) by explicitly specifying a set of accepted status + codes. Can be any of the following entities: + - an integer type specifying an exact failed status code to + accept + - a list or a tuple of integer types specifying a list of + failed status codes to accept + - a callable accepting an actual failed status code and + returning True if it should be accepted + Note that this argument does not affect success status codes (2xx) + which are always accepted. + """ + success = False try_count = 0 while success is False: try: - res = self._download_webpage_handle(url_or_request, video_id, note, errnote, fatal, encoding=encoding, data=data, headers=headers, query=query) + res = self._download_webpage_handle( + url_or_request, video_id, note, errnote, fatal, + encoding=encoding, data=data, headers=headers, query=query, + expected_status=expected_status) success = True except compat_http_client.IncompleteRead as e: try_count += 1 @@ -642,19 +782,43 @@ class InfoExtractor(object): content, _ = res return content - def _download_xml(self, url_or_request, video_id, - note='Downloading XML', errnote='Unable to download XML', - transform_source=None, fatal=True, encoding=None, - data=None, headers={}, query={}): - """Return the xml as an xml.etree.ElementTree.Element""" - xml_string = self._download_webpage( + def _download_xml_handle( + self, url_or_request, video_id, note='Downloading XML', + errnote='Unable to download XML', transform_source=None, + fatal=True, encoding=None, data=None, headers={}, query={}, + expected_status=None): + """ + Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle). + + See _download_webpage docstring for arguments specification. + """ + res = self._download_webpage_handle( url_or_request, video_id, note, errnote, fatal=fatal, - encoding=encoding, data=data, headers=headers, query=query) - if xml_string is False: - return xml_string + encoding=encoding, data=data, headers=headers, query=query, + expected_status=expected_status) + if res is False: + return res + xml_string, urlh = res return self._parse_xml( xml_string, video_id, transform_source=transform_source, - fatal=fatal) + fatal=fatal), urlh + + def _download_xml( + self, url_or_request, video_id, + note='Downloading XML', errnote='Unable to download XML', + transform_source=None, fatal=True, encoding=None, + data=None, headers={}, query={}, expected_status=None): + """ + Return the xml as an xml.etree.ElementTree.Element. + + See _download_webpage docstring for arguments specification. + """ + res = self._download_xml_handle( + url_or_request, video_id, note=note, errnote=errnote, + transform_source=transform_source, fatal=fatal, encoding=encoding, + data=data, headers=headers, query=query, + expected_status=expected_status) + return res if res is False else res[0] def _parse_xml(self, xml_string, video_id, transform_source=None, fatal=True): if transform_source: @@ -668,18 +832,43 @@ class InfoExtractor(object): else: self.report_warning(errmsg + str(ve)) - def _download_json(self, url_or_request, video_id, - note='Downloading JSON metadata', - errnote='Unable to download JSON metadata', - transform_source=None, - fatal=True, encoding=None, data=None, headers={}, query={}): - json_string = self._download_webpage( + def _download_json_handle( + self, url_or_request, video_id, note='Downloading JSON metadata', + errnote='Unable to download JSON metadata', transform_source=None, + fatal=True, encoding=None, data=None, headers={}, query={}, + expected_status=None): + """ + Return a tuple (JSON object, URL handle). + + See _download_webpage docstring for arguments specification. + """ + res = self._download_webpage_handle( url_or_request, video_id, note, errnote, fatal=fatal, - encoding=encoding, data=data, headers=headers, query=query) - if (not fatal) and json_string is False: - return None + encoding=encoding, data=data, headers=headers, query=query, + expected_status=expected_status) + if res is False: + return res + json_string, urlh = res return self._parse_json( - json_string, video_id, transform_source=transform_source, fatal=fatal) + json_string, video_id, transform_source=transform_source, + fatal=fatal), urlh + + def _download_json( + self, url_or_request, video_id, note='Downloading JSON metadata', + errnote='Unable to download JSON metadata', transform_source=None, + fatal=True, encoding=None, data=None, headers={}, query={}, + expected_status=None): + """ + Return the JSON object as a dict. + + See _download_webpage docstring for arguments specification. + """ + res = self._download_json_handle( + url_or_request, video_id, note=note, errnote=errnote, + transform_source=transform_source, fatal=fatal, encoding=encoding, + data=data, headers=headers, query=query, + expected_status=expected_status) + return res if res is False else res[0] def _parse_json(self, json_string, video_id, transform_source=None, fatal=True): if transform_source: @@ -972,8 +1161,7 @@ class InfoExtractor(object): def _search_json_ld(self, html, video_id, expected_type=None, **kwargs): json_ld = self._search_regex( - r'(?s)]+type=(["\'])application/ld\+json\1[^>]*>(?P.+?)', - html, 'JSON-LD', group='json_ld', **kwargs) + JSON_LD_RE, html, 'JSON-LD', group='json_ld', **kwargs) default = kwargs.get('default', NO_DEFAULT) if not json_ld: return default if default is not NO_DEFAULT else {} @@ -994,13 +1182,47 @@ class InfoExtractor(object): if isinstance(json_ld, dict): json_ld = [json_ld] + INTERACTION_TYPE_MAP = { + 'CommentAction': 'comment', + 'AgreeAction': 'like', + 'DisagreeAction': 'dislike', + 'LikeAction': 'like', + 'DislikeAction': 'dislike', + 'ListenAction': 'view', + 'WatchAction': 'view', + 'ViewAction': 'view', + } + + def extract_interaction_statistic(e): + interaction_statistic = e.get('interactionStatistic') + if not isinstance(interaction_statistic, list): + return + for is_e in interaction_statistic: + if not isinstance(is_e, dict): + continue + if is_e.get('@type') != 'InteractionCounter': + continue + interaction_type = is_e.get('interactionType') + if not isinstance(interaction_type, compat_str): + continue + interaction_count = int_or_none(is_e.get('userInteractionCount')) + if interaction_count is None: + continue + count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1]) + if not count_kind: + continue + count_key = '%s_count' % count_kind + if info.get(count_key) is not None: + continue + info[count_key] = interaction_count + def extract_video_object(e): assert e['@type'] == 'VideoObject' info.update({ - 'url': e.get('contentUrl'), + 'url': url_or_none(e.get('contentUrl')), 'title': unescapeHTML(e.get('name')), 'description': unescapeHTML(e.get('description')), - 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'), + 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')), 'duration': parse_duration(e.get('duration')), 'timestamp': unified_timestamp(e.get('uploadDate')), 'filesize': float_or_none(e.get('contentSize')), @@ -1009,24 +1231,35 @@ class InfoExtractor(object): 'height': int_or_none(e.get('height')), 'view_count': int_or_none(e.get('interactionCount')), }) + extract_interaction_statistic(e) for e in json_ld: - if e.get('@context') == 'http://schema.org': + if isinstance(e.get('@context'), compat_str) and re.match(r'^https?://schema.org/?$', e.get('@context')): item_type = e.get('@type') if expected_type is not None and expected_type != item_type: return info if item_type in ('TVEpisode', 'Episode'): + episode_name = unescapeHTML(e.get('name')) info.update({ - 'episode': unescapeHTML(e.get('name')), + 'episode': episode_name, 'episode_number': int_or_none(e.get('episodeNumber')), 'description': unescapeHTML(e.get('description')), }) + if not info.get('title') and episode_name: + info['title'] = episode_name part_of_season = e.get('partOfSeason') if isinstance(part_of_season, dict) and part_of_season.get('@type') in ('TVSeason', 'Season', 'CreativeWorkSeason'): info['season_number'] = int_or_none(part_of_season.get('seasonNumber')) part_of_series = e.get('partOfSeries') or e.get('partOfTVSeries') if isinstance(part_of_series, dict) and part_of_series.get('@type') in ('TVSeries', 'Series', 'CreativeWorkSeries'): info['series'] = unescapeHTML(part_of_series.get('name')) + elif item_type == 'Movie': + info.update({ + 'title': unescapeHTML(e.get('name')), + 'description': unescapeHTML(e.get('description')), + 'duration': parse_duration(e.get('duration')), + 'timestamp': unified_timestamp(e.get('dateCreated')), + }) elif item_type in ('Article', 'NewsArticle'): info.update({ 'timestamp': parse_iso8601(e.get('datePublished')), @@ -1489,9 +1722,9 @@ class InfoExtractor(object): # However, this is not always respected, for example, [2] # contains EXT-X-STREAM-INF tag which references AUDIO # rendition group but does not have CODECS and despite - # referencing audio group an audio group, it represents - # a complete (with audio and video) format. So, for such cases - # we will ignore references to rendition groups and treat them + # referencing an audio group it represents a complete + # (with audio and video) format. So, for such cases we will + # ignore references to rendition groups and treat them # as complete formats. if audio_group_id and codecs and f.get('vcodec') != 'none': audio_group = groups.get(audio_group_id) @@ -1647,9 +1880,7 @@ class InfoExtractor(object): 'height': height, }) formats.extend(m3u8_formats) - continue - - if src_ext == 'f4m': + elif src_ext == 'f4m': f4m_url = src_url if not f4m_params: f4m_params = { @@ -1659,9 +1890,13 @@ class InfoExtractor(object): 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', fatal=False)) - continue - - if src_url.startswith('http') and self._is_valid_url(src, video_id): + elif src_ext == 'mpd': + formats.extend(self._extract_mpd_formats( + src_url, video_id, mpd_id='dash', fatal=False)) + elif re.search(r'\.ism/[Mm]anifest', src_url): + formats.extend(self._extract_ism_formats( + src_url, video_id, ism_id='mss', fatal=False)) + elif src_url.startswith('http') and self._is_valid_url(src, video_id): http_count += 1 formats.append({ 'url': src_url, @@ -1672,7 +1907,6 @@ class InfoExtractor(object): 'width': width, 'height': height, }) - continue return formats @@ -1692,22 +1926,24 @@ class InfoExtractor(object): }) return subtitles - def _extract_xspf_playlist(self, playlist_url, playlist_id, fatal=True): + def _extract_xspf_playlist(self, xspf_url, playlist_id, fatal=True): xspf = self._download_xml( - playlist_url, playlist_id, 'Downloading xpsf playlist', + xspf_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) + return self._parse_xspf( + xspf, playlist_id, xspf_url=xspf_url, + xspf_base_url=base_url(xspf_url)) - def _parse_xspf(self, playlist, playlist_id): + def _parse_xspf(self, xspf_doc, playlist_id, xspf_url=None, xspf_base_url=None): 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)): + for track in xspf_doc.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( @@ -1717,12 +1953,18 @@ class InfoExtractor(object): 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))] + formats = [] + for location in track.findall(xpath_with_ns('./xspf:location', NS_MAP)): + format_url = urljoin(xspf_base_url, location.text) + if not format_url: + continue + formats.append({ + 'url': format_url, + 'manifest_url': xspf_url, + '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))), + }) self._sort_formats(formats) entries.append({ @@ -1736,18 +1978,18 @@ class InfoExtractor(object): return entries def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}): - res = self._download_webpage_handle( + res = self._download_xml_handle( mpd_url, video_id, note=note or 'Downloading MPD manifest', errnote=errnote or 'Failed to download MPD manifest', fatal=fatal) if res is False: return [] - mpd, urlh = res + mpd_doc, urlh = res mpd_base_url = base_url(urlh.geturl()) return self._parse_mpd_formats( - compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, + mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url, formats_dict=formats_dict, mpd_url=mpd_url) def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None): @@ -1886,7 +2128,21 @@ class InfoExtractor(object): representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info) def prepare_template(template_name, identifiers): - t = representation_ms_info[template_name] + tmpl = representation_ms_info[template_name] + # First of, % characters outside $...$ templates + # must be escaped by doubling for proper processing + # by % operator string formatting used further (see + # https://github.com/rg3/youtube-dl/issues/16867). + t = '' + in_template = False + for c in tmpl: + t += c + if c == '$': + in_template = not in_template + elif c == '%' and not in_template: + t += c + # Next, $...$ templates are translated to their + # %(...) counterparts to be used with % operator t = t.replace('$RepresentationID$', representation_id) t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t) t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t) @@ -2021,17 +2277,16 @@ class InfoExtractor(object): return formats def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True): - res = self._download_webpage_handle( + res = self._download_xml_handle( ism_url, video_id, note=note or 'Downloading ISM manifest', errnote=errnote or 'Failed to download ISM manifest', fatal=fatal) if res is False: return [] - ism, urlh = res + ism_doc, urlh = res - return self._parse_ism_formats( - compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id) + return self._parse_ism_formats(ism_doc, urlh.geturl(), ism_id) def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None): """ @@ -2129,8 +2384,8 @@ class InfoExtractor(object): return formats def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None, preference=None): - def absolute_url(video_url): - return compat_urlparse.urljoin(base_url, video_url) + def absolute_url(item_url): + return urljoin(base_url, item_url) def parse_content_type(content_type): if not content_type: @@ -2187,7 +2442,7 @@ class InfoExtractor(object): if src: _, formats = _media_formats(src, media_type) media_info['formats'].extend(formats) - media_info['thumbnail'] = media_attributes.get('poster') + media_info['thumbnail'] = absolute_url(media_attributes.get('poster')) if media_content: for source_tag in re.findall(r']+>', media_content): source_attributes = extract_attributes(source_tag) @@ -2218,6 +2473,8 @@ class InfoExtractor(object): media_info['subtitles'].setdefault(lang, []).append({ 'url': absolute_url(src), }) + for f in media_info['formats']: + f.setdefault('http_headers', {})['Referer'] = base_url if media_info['formats'] or media_info['subtitles']: entries.append(media_info) return entries @@ -2248,9 +2505,10 @@ class InfoExtractor(object): def _extract_wowza_formats(self, url, video_id, m3u8_entry_protocol='m3u8_native', skip_protocols=[]): query = compat_urlparse.urlparse(url).query url = re.sub(r'/(?:manifest|playlist|jwplayer)\.(?:m3u8|f4m|mpd|smil)', '', url) - url_base = self._search_regex( - r'(?:(?:https?|rtmp|rtsp):)?(//[^?]+)', url, 'format url') - http_base_url = '%s:%s' % ('http', url_base) + mobj = re.search( + r'(?:(?:http|rtmp|rtsp)(?Ps)?:)?(?P//[^?]+)', url) + url_base = mobj.group('url') + http_base_url = '%s%s:%s' % ('http', mobj.group('s') or '', url_base) formats = [] def manifest_url(manifest): @@ -2350,7 +2608,10 @@ class InfoExtractor(object): for track in tracks: if not isinstance(track, dict): continue - if track.get('kind') != 'captions': + track_kind = track.get('kind') + if not track_kind or not isinstance(track_kind, compat_str): + continue + if track_kind.lower() not in ('captions', 'subtitles'): continue track_url = urljoin(base_url, track.get('file')) if not track_url: