X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/32ad8acade51269177cd01c5aa3c7169557eeac8..4d073ced87f1fff7fd6fdfd82462444769b9f90b:/youtube_dl/extractor/common.py diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index e2d9f52..a2548db 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -174,6 +174,8 @@ 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. @@ -301,8 +303,9 @@ 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", "description" and "id" attributes - with the same semantics as videos (see above). + Additionally, playlists can have "id", "title", "description", "uploader", + "uploader_id", "uploader_url" attributes with the same semantics as videos + (see above). _type "multi_video" indicates that there are multiple videos that @@ -336,15 +339,17 @@ 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. @@ -355,6 +360,7 @@ 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): @@ -389,12 +395,15 @@ 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. @@ -405,28 +414,82 @@ 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.""" @@ -494,6 +557,16 @@ class InfoExtractor(object): self.to_screen('%s' % (note,)) else: self.to_screen('%s: %s' % (video_id, note)) + + # Some sites check X-Forwarded-For HTTP header in order to figure out + # the origin of the client behind proxy. This allows bypassing geo + # restriction by faking this header's value to IP that belongs to some + # geo unrestricted country. We will do so once we encounter any + # geo restriction error. + if self._x_forwarded_for_ip: + if 'X-Forwarded-For' not in headers: + headers['X-Forwarded-For'] = self._x_forwarded_for_ip + if isinstance(url_or_request, compat_urllib_request.Request): url_or_request = update_Request( url_or_request, data=data, headers=headers, query=query) @@ -523,15 +596,6 @@ class InfoExtractor(object): if isinstance(url_or_request, (compat_str, str)): url_or_request = url_or_request.partition('#')[0] - # Some sites check X-Forwarded-For HTTP header in order to figure out - # the origin of the client behind proxy. This allows bypassing geo - # restriction by faking this header's value to IP that belongs to some - # geo unrestricted country. We will do so once we encounter any - # geo restriction error. - if self._x_forwarded_for_ip: - if 'X-Forwarded-For' not in headers: - headers['X-Forwarded-For'] = self._x_forwarded_for_ip - urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query) if urlh is False: assert not fatal @@ -592,19 +656,11 @@ class InfoExtractor(object): if not encoding: encoding = self._guess_encoding_from_content(content_type, webpage_bytes) if self._downloader.params.get('dump_intermediate_pages', False): - try: - url = url_or_request.get_full_url() - except AttributeError: - url = url_or_request - self.to_screen('Dumping request to ' + url) + self.to_screen('Dumping request to ' + urlh.geturl()) dump = base64.b64encode(webpage_bytes).decode('ascii') self._downloader.to_screen(dump) if self._downloader.params.get('write_pages', False): - try: - url = url_or_request.get_full_url() - except AttributeError: - url = url_or_request - basen = '%s_%s' % (video_id, url) + basen = '%s_%s' % (video_id, urlh.geturl()) if len(basen) > 240: h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest() basen = basen[:240 - len(h)] + h @@ -648,19 +704,31 @@ class InfoExtractor(object): content, _ = res return content + 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={}): + """Return a tuple (xml as an xml.etree.ElementTree.Element, URL handle)""" + res = self._download_webpage_handle( + url_or_request, video_id, note, errnote, fatal=fatal, + encoding=encoding, data=data, headers=headers, query=query) + if res is False: + return res + xml_string, urlh = res + return self._parse_xml( + xml_string, video_id, transform_source=transform_source, + 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={}): """Return the xml as an xml.etree.ElementTree.Element""" - xml_string = self._download_webpage( - 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 - return self._parse_xml( - xml_string, video_id, transform_source=transform_source, - fatal=fatal) + 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) + 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: @@ -674,18 +742,30 @@ 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={}): + """Return a tuple (JSON object, URL handle)""" + 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 + 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={}): + 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) + 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: @@ -1000,6 +1080,40 @@ 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({ @@ -1015,9 +1129,10 @@ 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 @@ -1033,7 +1148,7 @@ class InfoExtractor(object): 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 == 'Article': + elif item_type in ('Article', 'NewsArticle'): info.update({ 'timestamp': parse_iso8601(e.get('datePublished')), 'title': unescapeHTML(e.get('headline')), @@ -1356,6 +1471,9 @@ class InfoExtractor(object): if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access return [] + if re.search(r'#EXT-X-SESSION-KEY:.*?URI="skd://', m3u8_doc): # Apple FairPlay + return [] + formats = [] format_url = lambda u: ( @@ -1695,22 +1813,24 @@ 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( @@ -1720,12 +1840,18 @@ 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({ @@ -1739,18 +1865,18 @@ 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): @@ -1883,6 +2009,7 @@ class InfoExtractor(object): 'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None, 'format_note': 'DASH %s' % content_type, 'filesize': filesize, + 'container': mimetype2ext(mime_type) + '_dash', } f.update(parse_codecs(representation_attrib.get('codecs'))) representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info) @@ -1980,6 +2107,22 @@ class InfoExtractor(object): }) segment_index += 1 representation_ms_info['fragments'] = fragments + elif 'segment_urls' in representation_ms_info: + # Segment URLs with no SegmentTimeline + # Example: https://www.seznam.cz/zpravy/clanek/cesko-zasahne-vitr-o-sile-vichrice-muze-byt-i-zivotu-nebezpecny-39091 + # https://github.com/rg3/youtube-dl/pull/14844 + fragments = [] + segment_duration = float_or_none( + representation_ms_info['segment_duration'], + representation_ms_info['timescale']) if 'segment_duration' in representation_ms_info else None + for segment_url in representation_ms_info['segment_urls']: + fragment = { + location_key(segment_url): segment_url, + } + if segment_duration: + fragment['duration'] = segment_duration + fragments.append(fragment) + representation_ms_info['fragments'] = fragments # NB: MPD manifest may contain direct URLs to unfragmented media. # No fragments key is present in this case. if 'fragments' in representation_ms_info: @@ -1994,32 +2137,29 @@ class InfoExtractor(object): f['url'] = initialization_url f['fragments'].append({location_key(initialization_url): initialization_url}) f['fragments'].extend(representation_ms_info['fragments']) - try: - existing_format = next( - fo for fo in formats - if fo['format_id'] == representation_id) - except StopIteration: - full_info = formats_dict.get(representation_id, {}).copy() - full_info.update(f) - formats.append(full_info) - else: - existing_format.update(f) + # According to [1, 5.3.5.2, Table 7, page 35] @id of Representation + # is not necessarily unique within a Period thus formats with + # the same `format_id` are quite possible. There are numerous examples + # of such manifests (see https://github.com/rg3/youtube-dl/issues/15111, + # https://github.com/rg3/youtube-dl/issues/13919) + full_info = formats_dict.get(representation_id, {}).copy() + full_info.update(f) + formats.append(full_info) else: self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type) 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): """ @@ -2043,7 +2183,7 @@ class InfoExtractor(object): stream_timescale = int_or_none(stream.get('TimeScale')) or timescale stream_name = stream.get('Name') for track in stream.findall('QualityLevel'): - fourcc = track.get('FourCC') + fourcc = track.get('FourCC', 'AACL' if track.get('AudioTag') == '255' else None) # TODO: add support for WVC1 and WMAP if fourcc not in ('H264', 'AVC1', 'AACL'): self.report_warning('%s is not a supported codec' % fourcc) @@ -2117,8 +2257,8 @@ 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: @@ -2175,7 +2315,7 @@ 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) @@ -2236,9 +2376,10 @@ 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): @@ -2338,7 +2479,10 @@ 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: @@ -2392,7 +2536,7 @@ class InfoExtractor(object): formats.extend(self._extract_m3u8_formats( source_url, video_id, 'mp4', entry_protocol='m3u8_native', m3u8_id=m3u8_id, fatal=False)) - elif ext == 'mpd': + elif source_type == 'dash' or ext == 'mpd': formats.extend(self._extract_mpd_formats( source_url, video_id, mpd_id=mpd_id, fatal=False)) elif ext == 'smil':