ansol/youtube-dl
7
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-12-19 11:34:04 +00:00
Code Issues Projects Releases Wiki Activity

[amp] imporove thumbnail and subtitle extraction

Browse Source
This commit is contained in:
Remita Amine 2017-05-02 00:06:19 +01:00
parent c6fe5a7e12
commit 7fc875195f
1 changed files with 13 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
youtube_dl/extractor/amp.py
View File

@ -34,9 +34,12 @@ class AMPIE(InfoExtractor):
if isinstance(media_thumbnail, dict): if isinstance(media_thumbnail, dict):
media_thumbnail = [media_thumbnail] media_thumbnail = [media_thumbnail]
for thumbnail_data in media_thumbnail: for thumbnail_data in media_thumbnail:
thumbnail = thumbnail_data['@attributes'] thumbnail = thumbnail_data.get('@attributes', {})
thumbnail_url = thumbnail.get('url')
if not thumbnail_url:
continue
thumbnails.append({ thumbnails.append({
'url': self._proto_relative_url(thumbnail['url'], 'http:'), 'url': self._proto_relative_url(thumbnail_url, 'http:'),
'width': int_or_none(thumbnail.get('width')), 'width': int_or_none(thumbnail.get('width')),
'height': int_or_none(thumbnail.get('height')), 'height': int_or_none(thumbnail.get('height')),
}) })
@ -47,9 +50,14 @@ class AMPIE(InfoExtractor):
if isinstance(media_subtitle, dict): if isinstance(media_subtitle, dict):
media_subtitle = [media_subtitle] media_subtitle = [media_subtitle]
for subtitle_data in media_subtitle: for subtitle_data in media_subtitle:
subtitle = subtitle_data['@attributes'] subtitle = subtitle_data.get('@attributes', {})
lang = subtitle.get('lang') or 'en' subtitle_href = subtitle.get('href')
subtitles[lang] = [{'url': subtitle['href']}] if not subtitle_href:
continue
subtitles.setdefault(subtitle.get('lang') or 'en', []).append({
'url': subtitle_href,
'ext': mimetype2ext(subtitle.get('type')) or determine_ext(subtitle_href),
})
formats = [] formats = []
media_content = get_media_node('content') media_content = get_media_node('content')