From 9be31e771cd9481ea690c01eed398645deadc1de Mon Sep 17 00:00:00 2001 From: Yen Chi Hsuan Date: Tue, 11 Jul 2017 15:48:34 +0800 Subject: [PATCH] [twitter] Support HLS streams in vmap URLs --- ChangeLog | 1 + youtube_dl/extractor/twitter.py | 23 ++++++++++++++--------- 2 files changed, 15 insertions(+), 9 deletions(-) diff --git a/ChangeLog b/ChangeLog index ffb4b69945..8e64511441 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,6 +1,7 @@ version Extractors ++ [twitter] Support HLS streams in vmap URLs + [periscope] Support pscp.tv URLs in embedded frames * [niconico] Fix authentication error handling (#12486) * [giantbomb] Extract m3u8 formats (#13626) diff --git a/youtube_dl/extractor/twitter.py b/youtube_dl/extractor/twitter.py index e4bc7e012c..89eabe77e0 100644 --- a/youtube_dl/extractor/twitter.py +++ b/youtube_dl/extractor/twitter.py @@ -20,9 +20,16 @@ class TwitterBaseIE(InfoExtractor): - def _get_vmap_video_url(self, vmap_url, video_id): + def _extract_formats_from_vmap_url(self, vmap_url, video_id): vmap_data = self._download_xml(vmap_url, video_id) - return xpath_text(vmap_data, './/MediaFile').strip() + video_url = xpath_text(vmap_data, './/MediaFile').strip() + if determine_ext(video_url) == 'm3u8': + return self._extract_m3u8_formats( + video_url, video_id, ext='mp4', m3u8_id='hls', + entry_protocol='m3u8_native') + return [{ + 'url': video_url, + }] @staticmethod def _search_dimensions_in_video_url(a_format, video_url): @@ -197,9 +204,8 @@ def _real_extract(self, url): vmap_url = config.get('vmapUrl') or config.get('vmap_url') if vmap_url: - formats.append({ - 'url': self._get_vmap_video_url(vmap_url, video_id), - }) + formats.extend( + self._extract_formats_from_vmap_url(vmap_url, video_id)) media_info = None @@ -449,7 +455,7 @@ def _real_extract(self, url): vmap_url = self._html_search_meta( 'twitter:amplify:vmap', webpage, 'vmap url') - video_url = self._get_vmap_video_url(vmap_url, video_id) + formats = self._extract_formats_from_vmap_url(vmap_url, video_id) thumbnails = [] thumbnail = self._html_search_meta( @@ -471,11 +477,10 @@ def _find_dimension(target): }) video_w, video_h = _find_dimension('player') - formats = [{ - 'url': video_url, + formats[0].update({ 'width': video_w, 'height': video_h, - }] + }) return { 'id': video_id,