X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/848723ea972c09f28787db91d8c06e98a274ab89..a5a28b1eabe0533f26d23c4157c97144215b22b7:/youtube_dl/extractor/theplatform.py?ds=inline diff --git a/youtube_dl/extractor/theplatform.py b/youtube_dl/extractor/theplatform.py index d607023..b1a985f 100644 --- a/youtube_dl/extractor/theplatform.py +++ b/youtube_dl/extractor/theplatform.py @@ -1,104 +1,397 @@ +# coding: utf-8 +from __future__ import unicode_literals + import re -import json +import time +import hmac +import binascii +import hashlib + -from .common import InfoExtractor +from .once import OnceIE +from .adobepass import AdobePassIE +from ..compat import ( + compat_parse_qs, + compat_urllib_parse_urlparse, +) from ..utils import ( + determine_ext, ExtractorError, + float_or_none, + int_or_none, + sanitized_Request, + unsmuggle_url, + update_url_query, xpath_with_ns, + mimetype2ext, + find_xpath_attr, ) -_x = lambda p: xpath_with_ns(p, {'smil': 'http://www.w3.org/2005/SMIL21/Language'}) +default_ns = 'http://www.w3.org/2005/SMIL21/Language' +_x = lambda p: xpath_with_ns(p, {'smil': default_ns}) + + +class ThePlatformBaseIE(OnceIE): + def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'): + meta = self._download_xml( + smil_url, video_id, note=note, query={'format': 'SMIL'}, + headers=self.geo_verification_headers()) + error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src') + if error_element is not None and error_element.attrib['src'].startswith( + 'http://link.theplatform.com/s/errorFiles/Unavailable.'): + raise ExtractorError(error_element.attrib['abstract'], expected=True) + + smil_formats = self._parse_smil_formats( + meta, smil_url, video_id, namespace=default_ns, + # the parameters are from syfy.com, other sites may use others, + # they also work for nbc.com + f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'}, + transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src)) + + formats = [] + for _format in smil_formats: + if OnceIE.suitable(_format['url']): + formats.extend(self._extract_once_formats(_format['url'])) + else: + media_url = _format['url'] + if determine_ext(media_url) == 'm3u8': + hdnea2 = self._get_cookies(media_url).get('hdnea2') + if hdnea2: + _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value}) + + formats.append(_format) + + subtitles = self._parse_smil_subtitles(meta, default_ns) + + return formats, subtitles + + def _download_theplatform_metadata(self, path, video_id): + info_url = 'http://link.theplatform.com/s/%s?format=preview' % path + return self._download_json(info_url, video_id) + + def _parse_theplatform_metadata(self, info): + subtitles = {} + captions = info.get('captions') + if isinstance(captions, list): + for caption in captions: + lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type') + subtitles.setdefault(lang, []).append({ + 'ext': mimetype2ext(mime), + 'url': src, + }) + + duration = info.get('duration') + tp_chapters = info.get('chapters', []) + chapters = [] + if tp_chapters: + def _add_chapter(start_time, end_time): + start_time = float_or_none(start_time, 1000) + end_time = float_or_none(end_time, 1000) + if start_time is None or end_time is None: + return + chapters.append({ + 'start_time': start_time, + 'end_time': end_time, + }) + + for chapter in tp_chapters[:-1]: + _add_chapter(chapter.get('startTime'), chapter.get('endTime')) + _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration) + + return { + 'title': info['title'], + 'subtitles': subtitles, + 'description': info['description'], + 'thumbnail': info['defaultThumbnailUrl'], + 'duration': float_or_none(duration, 1000), + 'timestamp': int_or_none(info.get('pubDate'), 1000) or None, + 'uploader': info.get('billingCode'), + 'chapters': chapters, + } + + def _extract_theplatform_metadata(self, path, video_id): + info = self._download_theplatform_metadata(path, video_id) + return self._parse_theplatform_metadata(info) -class ThePlatformIE(InfoExtractor): +class ThePlatformIE(ThePlatformBaseIE, AdobePassIE): _VALID_URL = r'''(?x) - (?:https?://(?:link|player)\.theplatform\.com/[sp]/[^/]+/ - (?P[^/\?]+/(?:swf|config)/select/)? + (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P[^/]+)/ + (?:(?:(?:[^/]+/)+select/)?(?Pmedia/(?:guid/\d+/)?)?|(?P(?:[^/\?]+/(?:swf|config)|onsite)/select/))? |theplatform:)(?P[^/\?&]+)''' - _TEST = { + _TESTS = [{ # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/ - u'url': u'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true', - u'info_dict': { - u'id': u'e9I_cZgTgIPd', - u'ext': u'flv', - u'title': u'Blackberry\'s big, bold Z30', - u'description': u'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.', - u'duration': 247, + 'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true', + 'info_dict': { + 'id': 'e9I_cZgTgIPd', + 'ext': 'flv', + 'title': 'Blackberry\'s big, bold Z30', + 'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.', + 'duration': 247, + 'timestamp': 1383239700, + 'upload_date': '20131031', + 'uploader': 'CBSI-NEW', }, - u'params': { + 'params': { # rtmp download - u'skip_download': True, + 'skip_download': True, }, - } - - def _get_info(self, video_id, smil_url): - meta = self._download_xml(smil_url, video_id) - - try: - error_msg = next( - n.attrib['abstract'] - for n in meta.findall(_x('.//smil:ref')) - if n.attrib.get('title') == u'Geographic Restriction') - except StopIteration: - pass - else: - raise ExtractorError(error_msg, expected=True) - - info_url = 'http://link.theplatform.com/s/dJ5BDC/{0}?format=preview'.format(video_id) - info_json = self._download_webpage(info_url, video_id) - info = json.loads(info_json) - - head = meta.find(_x('smil:head')) - body = meta.find(_x('smil:body')) - - f4m_node = body.find(_x('smil:seq/smil:video')) - if f4m_node is not None: - formats = [{ - 'ext': 'flv', - # the parameters are from syfy.com, other sites may use others - 'url': f4m_node.attrib['src'] + '?g=UXWGVKRWHFSP&hdcore=3.0.3', - }] + 'skip': '404 Not Found', + }, { + # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/ + 'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT', + 'info_dict': { + 'id': '22d_qsQ6MIRT', + 'ext': 'flv', + 'description': 'md5:ac330c9258c04f9d7512cf26b9595409', + 'title': 'Tesla Model S: A second step towards a cleaner motoring future', + 'timestamp': 1426176191, + 'upload_date': '20150312', + 'uploader': 'CBSI-NEW', + }, + 'params': { + # rtmp download + 'skip_download': True, + } + }, { + 'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD', + 'info_dict': { + 'id': 'yMBg9E8KFxZD', + 'ext': 'mp4', + 'description': 'md5:644ad9188d655b742f942bf2e06b002d', + 'title': 'HIGHLIGHTS: USA bag first ever series Cup win', + 'uploader': 'EGSM', + } + }, { + 'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7', + 'only_matching': True, + }, { + 'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701', + 'md5': 'fb96bb3d85118930a5b055783a3bd992', + 'info_dict': { + 'id': 'tdy_or_siri_150701', + 'ext': 'mp4', + 'title': 'iPhone Siri’s sassy response to a math question has people talking', + 'description': 'md5:a565d1deadd5086f3331d57298ec6333', + 'duration': 83.0, + 'thumbnail': r're:^https?://.*\.jpg$', + 'timestamp': 1435752600, + 'upload_date': '20150701', + 'uploader': 'NBCU-NEWS', + }, + }, { + # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1 + # geo-restricted (US), HLS encrypted with AES-128 + 'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781', + 'only_matching': True, + }] + + @classmethod + def _extract_urls(cls, webpage): + m = re.search( + r'''(?x) + https?://player\.theplatform\.com/p/.+?)\2 + ''', webpage) + if m: + return [m.group('url')] + + # Are whitesapces ignored in URLs? + # https://github.com/rg3/youtube-dl/issues/12044 + matches = re.findall( + r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage) + if matches: + return [re.sub(r'\s', '', list(zip(*matches))[1][0])] + + @staticmethod + def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False): + flags = '10' if include_qs else '00' + expiration_date = '%x' % (int(time.time()) + life) + + def str_to_hex(str): + return binascii.b2a_hex(str.encode('ascii')).decode('ascii') + + def hex_to_bytes(hex): + return binascii.a2b_hex(hex.encode('ascii')) + + relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1) + clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path)) + checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest() + sig = flags + expiration_date + checksum + str_to_hex(sig_secret) + return '%s&sig=%s' % (url, sig) + + def _real_extract(self, url): + url, smuggled_data = unsmuggle_url(url, {}) + + mobj = re.match(self._VALID_URL, url) + provider_id = mobj.group('provider_id') + video_id = mobj.group('id') + + if not provider_id: + provider_id = 'dJ5BDC' + + path = provider_id + '/' + if mobj.group('media'): + path += mobj.group('media') + path += video_id + + qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query) + if 'guid' in qs_dict: + webpage = self._download_webpage(url, video_id) + scripts = re.findall(r']+src="([^"]+)"', webpage) + feed_id = None + # feed id usually locates in the last script. + # Seems there's no pattern for the interested script filename, so + # I try one by one + for script in reversed(scripts): + feed_script = self._download_webpage( + self._proto_relative_url(script, 'http:'), + video_id, 'Downloading feed script') + feed_id = self._search_regex( + r'defaultFeedId\s*:\s*"([^"]+)"', feed_script, + 'default feed id', default=None) + if feed_id is not None: + break + if feed_id is None: + raise ExtractorError('Unable to find feed id') + return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % ( + provider_id, feed_id, qs_dict['guid'][0])) + + if smuggled_data.get('force_smil_url', False): + smil_url = url + # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385) + elif '/guid/' in url: + headers = {} + source_url = smuggled_data.get('source_url') + if source_url: + headers['Referer'] = source_url + request = sanitized_Request(url, headers=headers) + webpage = self._download_webpage(request, video_id) + smil_url = self._search_regex( + r']+href=(["\'])(?P.+?)\1[^>]+type=["\']application/smil\+xml', + webpage, 'smil url', group='url') + path = self._search_regex( + r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path') + smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4' + elif mobj.group('config'): + config_url = url + '&form=json' + config_url = config_url.replace('swf/', 'config/') + config_url = config_url.replace('onsite/', 'onsite/config/') + config = self._download_json(config_url, video_id, 'Downloading config') + if 'releaseUrl' in config: + release_url = config['releaseUrl'] + else: + release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path + smil_url = release_url + '&formats=MPEG4&manifest=f4m' else: - base_url = head.find(_x('smil:meta')).attrib['base'] - switch = body.find(_x('smil:switch')) - formats = [] - for f in switch.findall(_x('smil:video')): - attr = f.attrib - width = int(attr['width']) - height = int(attr['height']) - vbr = int(attr['system-bitrate']) // 1000 - format_id = '%dx%d_%dk' % (width, height, vbr) - formats.append({ - 'format_id': format_id, - 'url': base_url, - 'play_path': 'mp4:' + attr['src'], - 'ext': 'flv', - 'width': width, - 'height': height, - 'vbr': vbr, - }) - self._sort_formats(formats) + smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path - return { + sig = smuggled_data.get('sig') + if sig: + smil_url = self._sign_url(smil_url, sig['key'], sig['secret']) + + formats, subtitles = self._extract_theplatform_smil(smil_url, video_id) + self._sort_formats(formats) + + ret = self._extract_theplatform_metadata(path, video_id) + combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles) + ret.update({ 'id': video_id, - 'title': info['title'], 'formats': formats, - 'description': info['description'], - 'thumbnail': info['defaultThumbnailUrl'], - 'duration': info['duration']//1000, - } - + 'subtitles': combined_subtitles, + }) + + return ret + + +class ThePlatformFeedIE(ThePlatformBaseIE): + _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s' + _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P[^/]+)/(?P[^?/]+)\?(?:[^&]+&)*(?Pby(?:Gui|I)d=(?P[\w-]+))' + _TESTS = [{ + # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207 + 'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207', + 'md5': '6e32495b5073ab414471b615c5ded394', + 'info_dict': { + 'id': 'n_hardball_5biden_140207', + 'ext': 'mp4', + 'title': 'The Biden factor: will Joe run in 2016?', + 'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.', + 'thumbnail': r're:^https?://.*\.jpg$', + 'upload_date': '20140208', + 'timestamp': 1391824260, + 'duration': 467.0, + 'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'], + 'uploader': 'NBCU-NEWS', + }, + }] + + def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None): + real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query) + entry = self._download_json(real_url, video_id)['entries'][0] + main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else None + + formats = [] + subtitles = {} + first_video_id = None + duration = None + asset_types = [] + for item in entry['media$content']: + smil_url = item['plfile$url'] + cur_video_id = ThePlatformIE._match_id(smil_url) + if first_video_id is None: + first_video_id = cur_video_id + duration = float_or_none(item.get('plfile$duration')) + for asset_type in item['plfile$assetTypes']: + if asset_type in asset_types: + continue + asset_types.append(asset_type) + query = { + 'mbr': 'true', + 'formats': item['plfile$format'], + 'assetTypes': asset_type, + } + if asset_type in asset_types_query: + query.update(asset_types_query[asset_type]) + cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query( + main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type) + formats.extend(cur_formats) + subtitles = self._merge_subtitles(subtitles, cur_subtitles) + + self._sort_formats(formats) + + thumbnails = [{ + 'url': thumbnail['plfile$url'], + 'width': int_or_none(thumbnail.get('plfile$width')), + 'height': int_or_none(thumbnail.get('plfile$height')), + } for thumbnail in entry.get('media$thumbnails', [])] + + timestamp = int_or_none(entry.get('media$availableDate'), scale=1000) + categories = [item['media$name'] for item in entry.get('media$categories', [])] + + ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id) + subtitles = self._merge_subtitles(subtitles, ret['subtitles']) + ret.update({ + 'id': video_id, + 'formats': formats, + 'subtitles': subtitles, + 'thumbnails': thumbnails, + 'duration': duration, + 'timestamp': timestamp, + 'categories': categories, + }) + if custom_fields: + ret.update(custom_fields(entry)) + + return ret + def _real_extract(self, url): mobj = re.match(self._VALID_URL, url) + video_id = mobj.group('id') - if mobj.group('config'): - config_url = url+ '&form=json' - config_url = config_url.replace('swf/', 'config/') - config_json = self._download_webpage(config_url, video_id, u'Downloading config') - config = json.loads(config_json) - smil_url = config['releaseUrl'] + '&format=SMIL&formats=MPEG4' - else: - smil_url = ('http://link.theplatform.com/s/dJ5BDC/{0}/meta.smil?' - 'format=smil&mbr=true'.format(video_id)) - return self._get_info(video_id, smil_url) + provider_id = mobj.group('provider_id') + feed_id = mobj.group('feed_id') + filter_query = mobj.group('filter') + + return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)