Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-02-24 08:39:59 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Merge pull request #158 from galgeek/aitfive-1668-soundcoud

Browse Source 
capture soundcloud user page before capturing tracks
This commit is contained in:
Noah Levitt 2019-08-15 15:46:55 -07:00 committed by GitHub
commit 16f886259d
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
brozzler/ydl.py
View File

@ -163,13 +163,21 @@ def _build_youtube_dl(worker, destdir, site):
ie_result['entries_no_dl'] = list(ie_result['entries']) ie_result['entries_no_dl'] = list(ie_result['entries'])
ie_result['entries'] = [] ie_result['entries'] = []
self.logger.info( self.logger.info(
'not downoading %s videos from this youtube ' 'not downloading %s videos from this youtube '
'playlist because we expect to capture them from ' 'playlist because we expect to capture them from '
'individual watch pages', 'individual watch pages',
len(ie_result['entries_no_dl'])) len(ie_result['entries_no_dl']))
elif ie.IE_NAME == 'soundcloud:user':
ie_result['entries_no_dl'] = list(ie_result['entries'])
ie_result['entries'] = []
self.logger.info(
'not downloading %s tracks from this soundcloud '
'user page because we expect to capture them from '
'individual track pages',
len(ie_result['entries_no_dl']))
else: else:
self.logger.info( self.logger.info(
'extractor %r found a video in %s', ie.IE_NAME, url) 'extractor %r found a download in %s', ie.IE_NAME, url)
def _push_stitched_up_vid_to_warcprox(self, site, info_dict, ctx): def _push_stitched_up_vid_to_warcprox(self, site, info_dict, ctx):
# XXX Don't know how to get the right content-type. Youtube-dl # XXX Don't know how to get the right content-type. Youtube-dl