X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/3477c644417600d9ec8f8d2a44f82da0a4b15eb5..aebf9c8f14c9f5d45f8c6aa523ee135fdfd421b2:/youtube_dl/extractor/kaltura.py diff --git a/youtube_dl/extractor/kaltura.py b/youtube_dl/extractor/kaltura.py index 0dcd6cd..ccbc39c 100644 --- a/youtube_dl/extractor/kaltura.py +++ b/youtube_dl/extractor/kaltura.py @@ -2,12 +2,18 @@ from __future__ import unicode_literals import re +import base64 from .common import InfoExtractor -from ..compat import compat_urllib_parse +from ..compat import ( + compat_urllib_parse, + compat_urlparse, +) from ..utils import ( + clean_html, ExtractorError, int_or_none, + unsmuggle_url, ) @@ -39,7 +45,7 @@ class KalturaIE(InfoExtractor): 'info_dict': { 'id': '1_1jc2y3e4', 'ext': 'mp4', - 'title': 'Track 4', + 'title': 'Straight from the Heart', 'upload_date': '20131219', 'uploader_id': 'mlundberg@wolfgangsvault.com', 'description': 'The Allman Brothers Band, 12/16/1981', @@ -109,43 +115,65 @@ class KalturaIE(InfoExtractor): 'version': '-1', }, { - 'action': 'getContextData', - 'contextDataParams:objectType': 'KalturaEntryContextDataParams', - 'contextDataParams:referrer': 'http://www.kaltura.com/', - 'contextDataParams:streamerType': 'http', + 'action': 'getbyentryid', 'entryId': video_id, - 'service': 'baseentry', + 'service': 'flavorAsset', }, ] return self._kaltura_api_call( video_id, actions, note='Downloading video info JSON') def _real_extract(self, url): + url, smuggled_data = unsmuggle_url(url, {}) + mobj = re.match(self._VALID_URL, url) partner_id = mobj.group('partner_id_s') or mobj.group('partner_id') or mobj.group('partner_id_html5') entry_id = mobj.group('id_s') or mobj.group('id') or mobj.group('id_html5') - info, source_data = self._get_video_info(entry_id, partner_id) - - formats = [{ - 'format_id': '%(fileExt)s-%(bitrate)s' % f, - 'ext': f['fileExt'], - 'tbr': f['bitrate'], - 'fps': f.get('frameRate'), - 'filesize_approx': int_or_none(f.get('size'), invscale=1024), - 'container': f.get('containerFormat'), - 'vcodec': f.get('videoCodecId'), - 'height': f.get('height'), - 'width': f.get('width'), - 'url': '%s/flavorId/%s' % (info['dataUrl'], f['id']), - } for f in source_data['flavorAssets']] + info, flavor_assets = self._get_video_info(entry_id, partner_id) + + source_url = smuggled_data.get('source_url') + if source_url: + referrer = base64.b64encode( + '://'.join(compat_urlparse.urlparse(source_url)[:2]) + .encode('utf-8')).decode('utf-8') + else: + referrer = None + + formats = [] + for f in flavor_assets: + # Continue if asset is not ready + if f['status'] != 2: + continue + video_url = '%s/flavorId/%s' % (info['dataUrl'], f['id']) + if referrer: + video_url += '?referrer=%s' % referrer + formats.append({ + 'format_id': '%(fileExt)s-%(bitrate)s' % f, + 'ext': f.get('fileExt'), + 'tbr': int_or_none(f['bitrate']), + 'fps': int_or_none(f.get('frameRate')), + 'filesize_approx': int_or_none(f.get('size'), invscale=1024), + 'container': f.get('containerFormat'), + 'vcodec': f.get('videoCodecId'), + 'height': int_or_none(f.get('height')), + 'width': int_or_none(f.get('width')), + 'url': video_url, + }) + m3u8_url = info['dataUrl'].replace('format/url', 'format/applehttp') + if referrer: + m3u8_url += '?referrer=%s' % referrer + formats.extend(self._extract_m3u8_formats( + m3u8_url, entry_id, 'mp4', 'm3u8_native', m3u8_id='hls', fatal=False)) + + self._check_formats(formats, entry_id) self._sort_formats(formats) return { 'id': entry_id, 'title': info['name'], 'formats': formats, - 'description': info.get('description'), + 'description': clean_html(info.get('description')), 'thumbnail': info.get('thumbnailUrl'), 'duration': info.get('duration'), 'timestamp': info.get('createdAt'),