diff options
| author | Sergey M․ <dstftw@gmail.com> | 2019-09-03 01:24:20 +0700 |
|---|---|---|
| committer | Sergey M․ <dstftw@gmail.com> | 2019-09-03 01:24:20 +0700 |
| commit | 31dbd054c801ec14c1ea29a2167b70c980f1d782 (patch) | |
| tree | e5e655e85dd9158e42ecd6ef04b40a8ab7d288c0 | |
| parent | 66d04c74e097c03e4d644d7292546884cbee3d2b (diff) | |
| download | youtube-dl-31dbd054c801ec14c1ea29a2167b70c980f1d782.tar.gz youtube-dl-31dbd054c801ec14c1ea29a2167b70c980f1d782.tar.xz youtube-dl-31dbd054c801ec14c1ea29a2167b70c980f1d782.zip | |
[platzi] Improve client data extraction (closes #22290)
| -rw-r--r-- | youtube_dl/extractor/platzi.py | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/youtube_dl/extractor/platzi.py b/youtube_dl/extractor/platzi.py index cd6b966c5..602207beb 100644 --- a/youtube_dl/extractor/platzi.py +++ b/youtube_dl/extractor/platzi.py @@ -107,7 +107,11 @@ class PlatziIE(PlatziBaseIE): data = self._parse_json( self._search_regex( - r'client_data\s*=\s*({.+?})\s*;', webpage, 'client data'), + # client_data may contain "};" so that we have to try more + # strict regex first + (r'client_data\s*=\s*({.+?})\s*;\s*\n', + r'client_data\s*=\s*({.+?})\s*;'), + webpage, 'client data'), lecture_id) material = data['initialState']['material'] |