X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/af014acd27e0b471d5903630847eabb26437b46c..39393b81acfaf4045fb7f20454a0226f0dc9142e:/youtube_dl/extractor/nexx.py diff --git a/youtube_dl/extractor/nexx.py b/youtube_dl/extractor/nexx.py index d0235fd..586c1b7 100644 --- a/youtube_dl/extractor/nexx.py +++ b/youtube_dl/extractor/nexx.py @@ -18,27 +18,30 @@ from ..utils import ( class NexxIE(InfoExtractor): - _VALID_URL = r'https?://api\.nexx(?:\.cloud|cdn\.com)/v3/(?P\d+)/videos/byid/(?P\d+)' + _VALID_URL = r'''(?x) + (?: + https?://api\.nexx(?:\.cloud|cdn\.com)/v3/(?P\d+)/videos/byid/| + nexx:(?:(?P\d+):)?| + https?://arc\.nexx\.cloud/api/video/ + ) + (?P\d+) + ''' _TESTS = [{ # movie 'url': 'https://api.nexx.cloud/v3/748/videos/byid/128907', - 'md5': '16746bfc28c42049492385c989b26c4a', + 'md5': '31899fd683de49ad46f4ee67e53e83fe', 'info_dict': { 'id': '128907', 'ext': 'mp4', 'title': 'Stiftung Warentest', 'alt_title': 'Wie ein Test abläuft', 'description': 'md5:d1ddb1ef63de721132abd38639cc2fd2', - 'release_year': 2013, 'creator': 'SPIEGEL TV', 'thumbnail': r're:^https?://.*\.jpg$', 'duration': 2509, 'timestamp': 1384264416, 'upload_date': '20131112', }, - 'params': { - 'format': 'bestvideo', - }, }, { # episode 'url': 'https://api.nexx.cloud/v3/741/videos/byid/247858', @@ -56,14 +59,59 @@ class NexxIE(InfoExtractor): 'season_number': 2, }, 'params': { - 'format': 'bestvideo', 'skip_download': True, }, + 'skip': 'HTTP Error 404: Not Found', + }, { + # does not work via arc + 'url': 'nexx:741:1269984', + 'md5': 'c714b5b238b2958dc8d5642addba6886', + 'info_dict': { + 'id': '1269984', + 'ext': 'mp4', + 'title': '1 TAG ohne KLO... wortwörtlich! 😑', + 'alt_title': '1 TAG ohne KLO... wortwörtlich! 😑', + 'thumbnail': r're:^https?://.*\.jpg$', + 'duration': 607, + 'timestamp': 1518614955, + 'upload_date': '20180214', + }, + }, { + # free cdn from http://www.spiegel.de/video/eifel-zoo-aufregung-um-ausgebrochene-raubtiere-video-99018031.html + 'url': 'nexx:747:1533779', + 'md5': '6bf6883912b82b7069fb86c2297e9893', + 'info_dict': { + 'id': '1533779', + 'ext': 'mp4', + 'title': 'Aufregung um ausgebrochene Raubtiere', + 'alt_title': 'Eifel-Zoo', + 'description': 'md5:f21375c91c74ad741dcb164c427999d2', + 'thumbnail': r're:^https?://.*\.jpg$', + 'duration': 111, + 'timestamp': 1527874460, + 'upload_date': '20180601', + }, }, { 'url': 'https://api.nexxcdn.com/v3/748/videos/byid/128907', 'only_matching': True, + }, { + 'url': 'nexx:748:128907', + 'only_matching': True, + }, { + 'url': 'nexx:128907', + 'only_matching': True, + }, { + 'url': 'https://arc.nexx.cloud/api/video/128907.json', + 'only_matching': True, }] + @staticmethod + def _extract_domain_id(webpage): + mobj = re.search( + r']+\bsrc=["\'](?:https?:)?//(?:require|arc)\.nexx(?:\.cloud|cdn\.com)/(?:sdk/)?(?P\d+)', + webpage) + return mobj.group('id') if mobj else None + @staticmethod def _extract_urls(webpage): # Reference: @@ -72,13 +120,10 @@ class NexxIE(InfoExtractor): entries = [] # JavaScript Integration - mobj = re.search( - r']+\bsrc=["\']https?://require\.nexx(?:\.cloud|cdn\.com)/(?P\d+)', - webpage) - if mobj: - domain_id = mobj.group('id') + domain_id = NexxIE._extract_domain_id(webpage) + if domain_id: for video_id in re.findall( - r'(?is)onPLAYReady.+?_play\.init\s*\(.+?\s*,\s*["\']?(\d+)', + r'(?is)onPLAYReady.+?_play\.(?:init|(?:control\.)?addPlayer)\s*\(.+?\s*,\s*["\']?(\d+)', webpage): entries.append( 'https://api.nexx.cloud/v3/%s/videos/byid/%s' @@ -110,104 +155,239 @@ class NexxIE(InfoExtractor): self._handle_error(result) return result['result'] - def _real_extract(self, url): - mobj = re.match(self._VALID_URL, url) - domain_id, video_id = mobj.group('domain_id', 'id') - - # Reverse engineered from JS code (see getDeviceID function) - device_id = '%d:%d:%d%d' % ( - random.randint(1, 4), int(time.time()), - random.randint(1e4, 99999), random.randint(1, 9)) - - result = self._call_api(domain_id, 'session/init', video_id, data={ - 'nxp_devh': device_id, - 'nxp_userh': '', - 'precid': '0', - 'playlicense': '0', - 'screenx': '1920', - 'screeny': '1080', - 'playerversion': '6.0.00', - 'gateway': 'html5', - 'adGateway': '', - 'explicitlanguage': 'en-US', - 'addTextTemplates': '1', - 'addDomainData': '1', - 'addAdModel': '1', - }, headers={ - 'X-Request-Enable-Auth-Fallback': '1', - }) - - cid = result['general']['cid'] - - # As described in [1] X-Request-Token generation algorithm is - # as follows: - # md5( operation + domain_id + domain_secret ) - # where domain_secret is a static value that will be given by nexx.tv - # as per [1]. Here is how this "secret" is generated (reversed - # from _play.api.init function, search for clienttoken). So it's - # actually not static and not that much of a secret. - # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf - secret = result['device']['clienttoken'][int(device_id[0]):] - secret = secret[0:len(secret) - int(device_id[-1])] - - op = 'byid' - - # Reversed from JS code for _play.api.call function (search for - # X-Request-Token) - request_token = hashlib.md5( - ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest() - - video = self._call_api( - domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={ - 'additionalfields': 'language,channel,actors,studio,licenseby,slug,subtitle,teaser,description', - 'addInteractionOptions': '1', - 'addStatusDetails': '1', - 'addStreamDetails': '1', - 'addCaptions': '1', - 'addScenes': '1', - 'addHotSpots': '1', - 'addBumpers': '1', - 'captionFormat': 'data', - }, headers={ - 'X-Request-CID': cid, - 'X-Request-Token': request_token, - }) - - general = video['general'] - title = general['title'] - + def _extract_free_formats(self, video, video_id): stream_data = video['streamdata'] - language = general.get('language_raw') or '' + cdn = stream_data['cdnType'] + assert cdn == 'free' + + hash = video['general']['hash'] + + ps = compat_str(stream_data['originalDomain']) + if stream_data['applyFolderHierarchy'] == 1: + s = ('%04d' % int(video_id))[::-1] + ps += '/%s/%s' % (s[0:2], s[2:4]) + ps += '/%s/%s_' % (video_id, hash) + + t = 'http://%s' + ps + fd = stream_data['azureFileDistribution'].split(',') + cdn_provider = stream_data['cdnProvider'] + + def p0(p): + return '_%s' % p if stream_data['applyAzureStructure'] == 1 else '' + + formats = [] + if cdn_provider == 'ak': + t += ',' + for i in fd: + p = i.split(':') + t += p[1] + p0(int(p[0])) + ',' + t += '.mp4.csmil/master.%s' + elif cdn_provider == 'ce': + k = t.split('/') + h = k.pop() + http_base = t = '/'.join(k) + http_base = http_base % stream_data['cdnPathHTTP'] + t += '/asset.ism/manifest.%s?dcp_ver=aos4&videostream=' + for i in fd: + p = i.split(':') + tbr = int(p[0]) + filename = '%s%s%s.mp4' % (h, p[1], p0(tbr)) + f = { + 'url': http_base + '/' + filename, + 'format_id': '%s-http-%d' % (cdn, tbr), + 'tbr': tbr, + } + width_height = p[1].split('x') + if len(width_height) == 2: + f.update({ + 'width': int_or_none(width_height[0]), + 'height': int_or_none(width_height[1]), + }) + formats.append(f) + a = filename + ':%s' % (tbr * 1000) + t += a + ',' + t = t[:-1] + '&audiostream=' + a.split(':')[0] + else: + assert False + + if cdn_provider == 'ce': + formats.extend(self._extract_mpd_formats( + t % (stream_data['cdnPathDASH'], 'mpd'), video_id, + mpd_id='%s-dash' % cdn, fatal=False)) + formats.extend(self._extract_m3u8_formats( + t % (stream_data['cdnPathHLS'], 'm3u8'), video_id, 'mp4', + entry_protocol='m3u8_native', m3u8_id='%s-hls' % cdn, fatal=False)) - # TODO: reverse more cdns and formats + return formats + def _extract_azure_formats(self, video, video_id): + stream_data = video['streamdata'] cdn = stream_data['cdnType'] assert cdn == 'azure' azure_locator = stream_data['azureLocator'] - AZURE_URL = 'http://nx-p%02d.akamaized.net/' - - for secure in ('s', ''): - cdn_shield = stream_data.get('cdnShieldHTTP%s' % secure.upper()) - if cdn_shield: - azure_base = 'http%s://%s' % (secure, cdn_shield) - break - else: - azure_base = AZURE_URL % int(stream_data['azureAccount'].replace('nexxplayplus', '')) - + def get_cdn_shield_base(shield_type='', static=False): + for secure in ('', 's'): + cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper())) + if cdn_shield: + return 'http%s://%s' % (secure, cdn_shield) + else: + if 'fb' in stream_data['azureAccount']: + prefix = 'df' if static else 'f' + else: + prefix = 'd' if static else 'p' + account = int(stream_data['azureAccount'].replace('nexxplayplus', '').replace('nexxplayfb', '')) + return 'http://nx-%s%02d.akamaized.net/' % (prefix, account) + + language = video['general'].get('language_raw') or '' + + azure_stream_base = get_cdn_shield_base() is_ml = ',' in language - azure_m3u8_url = '%s%s/%s_src%s.ism/Manifest(format=m3u8-aapl)' % ( - azure_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + azure_manifest_url = '%s%s/%s_src%s.ism/Manifest' % ( + azure_stream_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + '%s' protection_token = try_get( video, lambda x: x['protectiondata']['token'], compat_str) if protection_token: - azure_m3u8_url += '?hdnts=%s' % protection_token + azure_manifest_url += '?hdnts=%s' % protection_token formats = self._extract_m3u8_formats( - azure_m3u8_url, video_id, 'mp4', entry_protocol='m3u8_native', - m3u8_id='%s-hls' % cdn) + azure_manifest_url % '(format=m3u8-aapl)', + video_id, 'mp4', 'm3u8_native', + m3u8_id='%s-hls' % cdn, fatal=False) + formats.extend(self._extract_mpd_formats( + azure_manifest_url % '(format=mpd-time-csf)', + video_id, mpd_id='%s-dash' % cdn, fatal=False)) + formats.extend(self._extract_ism_formats( + azure_manifest_url % '', video_id, ism_id='%s-mss' % cdn, fatal=False)) + + azure_progressive_base = get_cdn_shield_base('Prog', True) + azure_file_distribution = stream_data.get('azureFileDistribution') + if azure_file_distribution: + fds = azure_file_distribution.split(',') + if fds: + for fd in fds: + ss = fd.split(':') + if len(ss) == 2: + tbr = int_or_none(ss[0]) + if tbr: + f = { + 'url': '%s%s/%s_src_%s_%d.mp4' % ( + azure_progressive_base, azure_locator, video_id, ss[1], tbr), + 'format_id': '%s-http-%d' % (cdn, tbr), + 'tbr': tbr, + } + width_height = ss[1].split('x') + if len(width_height) == 2: + f.update({ + 'width': int_or_none(width_height[0]), + 'height': int_or_none(width_height[1]), + }) + formats.append(f) + + return formats + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + domain_id = mobj.group('domain_id') or mobj.group('domain_id_s') + video_id = mobj.group('id') + + video = None + + def find_video(result): + if isinstance(result, dict): + return result + elif isinstance(result, list): + vid = int(video_id) + for v in result: + if try_get(v, lambda x: x['general']['ID'], int) == vid: + return v + return None + + response = self._download_json( + 'https://arc.nexx.cloud/api/video/%s.json' % video_id, + video_id, fatal=False) + if response and isinstance(response, dict): + result = response.get('result') + if result: + video = find_video(result) + + # not all videos work via arc, e.g. nexx:741:1269984 + if not video: + # Reverse engineered from JS code (see getDeviceID function) + device_id = '%d:%d:%d%d' % ( + random.randint(1, 4), int(time.time()), + random.randint(1e4, 99999), random.randint(1, 9)) + + result = self._call_api(domain_id, 'session/init', video_id, data={ + 'nxp_devh': device_id, + 'nxp_userh': '', + 'precid': '0', + 'playlicense': '0', + 'screenx': '1920', + 'screeny': '1080', + 'playerversion': '6.0.00', + 'gateway': 'html5', + 'adGateway': '', + 'explicitlanguage': 'en-US', + 'addTextTemplates': '1', + 'addDomainData': '1', + 'addAdModel': '1', + }, headers={ + 'X-Request-Enable-Auth-Fallback': '1', + }) + + cid = result['general']['cid'] + + # As described in [1] X-Request-Token generation algorithm is + # as follows: + # md5( operation + domain_id + domain_secret ) + # where domain_secret is a static value that will be given by nexx.tv + # as per [1]. Here is how this "secret" is generated (reversed + # from _play.api.init function, search for clienttoken). So it's + # actually not static and not that much of a secret. + # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf + secret = result['device']['clienttoken'][int(device_id[0]):] + secret = secret[0:len(secret) - int(device_id[-1])] + + op = 'byid' + + # Reversed from JS code for _play.api.call function (search for + # X-Request-Token) + request_token = hashlib.md5( + ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest() + + result = self._call_api( + domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={ + 'additionalfields': 'language,channel,actors,studio,licenseby,slug,subtitle,teaser,description', + 'addInteractionOptions': '1', + 'addStatusDetails': '1', + 'addStreamDetails': '1', + 'addCaptions': '1', + 'addScenes': '1', + 'addHotSpots': '1', + 'addBumpers': '1', + 'captionFormat': 'data', + }, headers={ + 'X-Request-CID': cid, + 'X-Request-Token': request_token, + }) + video = find_video(result) + + general = video['general'] + title = general['title'] + + cdn = video['streamdata']['cdnType'] + + if cdn == 'azure': + formats = self._extract_azure_formats(video, video_id) + elif cdn == 'free': + formats = self._extract_free_formats(video, video_id) + else: + # TODO: reverse more cdns + assert False + self._sort_formats(formats) return { @@ -230,8 +410,8 @@ class NexxIE(InfoExtractor): class NexxEmbedIE(InfoExtractor): - _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?P[^/?#&]+)' - _TEST = { + _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:video/)?(?P[^/?#&]+)' + _TESTS = [{ 'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1', 'md5': '16746bfc28c42049492385c989b26c4a', 'info_dict': { @@ -240,7 +420,6 @@ class NexxEmbedIE(InfoExtractor): 'title': 'Nervenkitzel Achterbahn', 'alt_title': 'Karussellbauer in Deutschland', 'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc', - 'release_year': 2005, 'creator': 'SPIEGEL TV', 'thumbnail': r're:^https?://.*\.jpg$', 'duration': 2761, @@ -251,7 +430,10 @@ class NexxEmbedIE(InfoExtractor): 'format': 'bestvideo', 'skip_download': True, }, - } + }, { + 'url': 'https://embed.nexx.cloud/11888/video/DSRTO7UVOX06S7', + 'only_matching': True, + }] @staticmethod def _extract_urls(webpage):