# coding: utf-8 from __future__ import unicode_literals import re from .common import InfoExtractor from ..utils import ( int_or_none, determine_ext, parse_age_limit, urlencode_postdata, ExtractorError, ) class GoIE(InfoExtractor): _BRANDS = { 'abc': '001', 'freeform': '002', 'watchdisneychannel': '004', 'watchdisneyjunior': '008', 'watchdisneyxd': '009', } _VALID_URL = r'https?://(?:(?P%s)\.)?go\.com/(?:[^/]+/)*(?:vdka(?P\w+)|season-\d+/\d+-(?P[^/?#]+))' % '|'.join(_BRANDS.keys()) _TESTS = [{ 'url': 'http://abc.go.com/shows/castle/video/most-recent/vdka0_g86w5onx', 'info_dict': { 'id': '0_g86w5onx', 'ext': 'mp4', 'title': 'Sneak Peek: Language Arts', 'description': 'md5:7dcdab3b2d17e5217c953256af964e9c', }, 'params': { # m3u8 download 'skip_download': True, }, }, { 'url': 'http://abc.go.com/shows/after-paradise/video/most-recent/vdka3335601', 'only_matching': True, }] def _real_extract(self, url): sub_domain, video_id, display_id = re.match(self._VALID_URL, url).groups() if not video_id: webpage = self._download_webpage(url, display_id) video_id = self._search_regex( # There may be inner quotes, e.g. data-video-id="'VDKA3609139'" # from http://freeform.go.com/shows/shadowhunters/episodes/season-2/1-this-guilty-blood r'data-video-id=["\']*VDKA(\w+)', webpage, 'video id') brand = self._BRANDS[sub_domain] video_data = self._download_json( 'http://api.contents.watchabc.go.com/vp2/ws/contents/3000/videos/%s/001/-1/-1/-1/%s/-1/-1.json' % (brand, video_id), video_id)['video'][0] title = video_data['title'] formats = [] for asset in video_data.get('assets', {}).get('asset', []): asset_url = asset.get('value') if not asset_url: continue format_id = asset.get('format') ext = determine_ext(asset_url) if ext == 'm3u8': video_type = video_data.get('type') if video_type == 'lf': entitlement = self._download_json( 'https://api.entitlement.watchabc.go.com/vp2/ws-secure/entitlement/2020/authorize.json', video_id, data=urlencode_postdata({ 'video_id': video_data['id'], 'video_type': video_type, 'brand': brand, 'device': '001', })) errors = entitlement.get('errors', {}).get('errors', []) if errors: error_message = ', '.join([error['message'] for error in errors]) raise ExtractorError('%s said: %s' % (self.IE_NAME, error_message), expected=True) asset_url += '?' + entitlement['uplynkData']['sessionKey'] formats.extend(self._extract_m3u8_formats( asset_url, video_id, 'mp4', m3u8_id=format_id or 'hls', fatal=False)) else: formats.append({ 'format_id': format_id, 'url': asset_url, 'ext': ext, }) self._sort_formats(formats) subtitles = {} for cc in video_data.get('closedcaption', {}).get('src', []): cc_url = cc.get('value') if not cc_url: continue ext = determine_ext(cc_url) if ext == 'xml': ext = 'ttml' subtitles.setdefault(cc.get('lang'), []).append({ 'url': cc_url, 'ext': ext, }) thumbnails = [] for thumbnail in video_data.get('thumbnails', {}).get('thumbnail', []): thumbnail_url = thumbnail.get('value') if not thumbnail_url: continue thumbnails.append({ 'url': thumbnail_url, 'width': int_or_none(thumbnail.get('width')), 'height': int_or_none(thumbnail.get('height')), }) return { 'id': video_id, 'title': title, 'description': video_data.get('longdescription') or video_data.get('description'), 'duration': int_or_none(video_data.get('duration', {}).get('value'), 1000), 'age_limit': parse_age_limit(video_data.get('tvrating', {}).get('rating')), 'episode_number': int_or_none(video_data.get('episodenumber')), 'series': video_data.get('show', {}).get('title'), 'season_number': int_or_none(video_data.get('season', {}).get('num')), 'thumbnails': thumbnails, 'formats': formats, 'subtitles': subtitles, }