Madeorsk/youtube-dl
1
0
Fork 0
mirror of https://codeberg.org/polarisfm/youtube-dl synced 2024-12-02 05:07:55 +01:00
Code Issues Releases Wiki Activity

[linkedin:learning] add support for subtitle

Browse Source
This commit is contained in:
Frank Wang 2019-04-25 22:19:39 +08:00
parent 58ef5e7881
commit 94c7e182e2
1 changed files with 24 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
youtube_dl/extractor/linkedin.py
View File

@ -2,6 +2,7 @@
from __future__ import unicode_literals from __future__ import unicode_literals
import re import re
from itertools import zip_longest
from .common import InfoExtractor from .common import InfoExtractor
from ..utils import ( from ..utils import (
@ -126,15 +127,37 @@ class LinkedInLearningIE(LinkedInLearningBaseIE):
self._sort_formats(formats, ('width', 'height', 'source_preference', 'tbr', 'abr')) self._sort_formats(formats, ('width', 'height', 'source_preference', 'tbr', 'abr'))
duration = int_or_none(video_data.get('durationInSeconds'))
subtitles = self.extract_subtitles(video_data['transcript'], duration)
return { return {
'id': self._get_video_id(video_data, course_slug, video_slug), 'id': self._get_video_id(video_data, course_slug, video_slug),
'title': title, 'title': title,
'formats': formats, 'formats': formats,
'subtitles': subtitles,
'thumbnail': video_data.get('defaultThumbnail'), 'thumbnail': video_data.get('defaultThumbnail'),
'timestamp': float_or_none(video_data.get('publishedOn'), 1000), 'timestamp': float_or_none(video_data.get('publishedOn'), 1000),
'duration': int_or_none(video_data.get('durationInSeconds')), 'duration': duration,
} }
def _get_subtitles(self, transcript, duration):
counter = 0
records = []
if transcript and 'lines' in transcript:
lines = transcript['lines']
for curr, next in zip_longest(lines, lines[1:], fillvalue={"transcriptStartAt": duration*1000, "caption": "THE END"}):
text = curr["caption"].strip()
if text:
counter += 1
show = curr["transcriptStartAt"]
hide = next["transcriptStartAt"]
show = "{:02d}:{:02d}:{:02d},{:03d}".format(show//3600000, show%3600000//60000, show%60000//1000, show%1000)
hide = "{:02d}:{:02d}:{:02d},{:03d}".format(hide//3600000, hide%3600000//60000, hide%60000//1000, hide%1000)
records.append('%s\r\n%s --> %s\r\n%s\r\n' % (counter, show, hide, text))
return {'en': [{'ext': 'srt', 'data': '\r\n'.join(records)}]}
else:
return {}
class LinkedInLearningCourseIE(LinkedInLearningBaseIE): class LinkedInLearningCourseIE(LinkedInLearningBaseIE):
IE_NAME = 'linkedin:learning:course' IE_NAME = 'linkedin:learning:course'