youtube-dl

commit 7fc875195fd74c235d6ca86afc642f683a485a4b
parent c6fe5a7e12a5ebbe137b889afb8aefc206d132af
Author: Remita Amine <remitamine@gmail.com>
Date:   Tue,  2 May 2017 00:06:19 +0100

[amp] imporove thumbnail and subtitle extraction

Diffstat:
M
youtube_dl/extractor/amp.py
 | 
18
+++++++++++++-----

1 file changed, 13 insertions(+), 5 deletions(-)
diff --git a/youtube_dl/extractor/amp.py b/youtube_dl/extractor/amp.py
@@ -34,9 +34,12 @@ class AMPIE(InfoExtractor):
             if isinstance(media_thumbnail, dict):
                 media_thumbnail = [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({
-                    'url': self._proto_relative_url(thumbnail['url'], 'http:'),
+                    'url': self._proto_relative_url(thumbnail_url, 'http:'),
                     'width': int_or_none(thumbnail.get('width')),
                     'height': int_or_none(thumbnail.get('height')),
                 })
@@ -47,9 +50,14 @@ class AMPIE(InfoExtractor):
             if isinstance(media_subtitle, dict):
                 media_subtitle = [media_subtitle]
             for subtitle_data in media_subtitle:
-                subtitle = subtitle_data['@attributes']
-                lang = subtitle.get('lang') or 'en'
-                subtitles[lang] = [{'url': subtitle['href']}]
+                subtitle = subtitle_data.get('@attributes', {})
+                subtitle_href = subtitle.get('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 = []
         media_content = get_media_node('content')