heckyel/yt-local
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix related video extraction sometimes failing

Browse Source 
Youtube added some pointless variation in variable names
This commit is contained in:
James Taylor 2020-04-10 13:09:38 -07:00
parent 3e09193eaf
commit 1224dd88a3
1 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
youtube/yt_data_extract/common.py
View File

@ -223,10 +223,14 @@ def extract_item_info(item, additional_info={}):
info['type'] = 'playlist' info['type'] = 'playlist'
elif primary_type == 'channel': elif primary_type == 'channel':
info['type'] = 'channel' info['type'] = 'channel'
elif type == 'videoWithContextRenderer': # stupid exception
info['type'] = 'video'
primary_type = 'video'
else: else:
info['type'] = 'unsupported' info['type'] = 'unsupported'
info['title'] = extract_str(item.get('title')) # videoWithContextRenderer changes it to 'headline' just to be annoying
info['title'] = extract_str(multi_get(item, 'title', 'headline'))
if primary_type != 'channel': if primary_type != 'channel':
info['author'] = extract_str(multi_get(item, 'longBylineText', 'shortBylineText', 'ownerText')) info['author'] = extract_str(multi_get(item, 'longBylineText', 'shortBylineText', 'ownerText'))
info['author_id'] = extract_str(multi_deep_get(item, info['author_id'] = extract_str(multi_deep_get(item,
@ -256,7 +260,10 @@ def extract_item_info(item, additional_info={}):
info['view_count'] = extract_int(item.get('viewCountText')) info['view_count'] = extract_int(item.get('viewCountText'))
# dig into accessibility data to get view_count for videos marked as recommended, and to get time_published # dig into accessibility data to get view_count for videos marked as recommended, and to get time_published
accessibility_label = deep_get(item, 'title', 'accessibility', 'accessibilityData', 'label', default='') accessibility_label = multi_deep_get(item,
['title', 'accessibility', 'accessibilityData', 'label'],
['headline', 'accessibility', 'accessibilityData', 'label'],
default='')
timestamp = re.search(r'(\d+ \w+ ago)', accessibility_label) timestamp = re.search(r'(\d+ \w+ ago)', accessibility_label)
if timestamp: if timestamp:
conservative_update(info, 'time_published', timestamp.group(1)) conservative_update(info, 'time_published', timestamp.group(1))
@ -333,6 +340,7 @@ _item_types = {
'videoRenderer', 'videoRenderer',
'compactVideoRenderer', 'compactVideoRenderer',
'compactAutoplayRenderer', 'compactAutoplayRenderer',
'videoWithContextRenderer',
'gridVideoRenderer', 'gridVideoRenderer',
'playlistVideoRenderer', 'playlistVideoRenderer',