[utils] add a function to clean podcast URLs

This commit is contained in:
Remita Amine 2021-01-04 00:51:55 +01:00
parent 0889eb33e0
commit e88c9ef62a
2 changed files with 22 additions and 0 deletions

View file

@ -21,6 +21,7 @@ from youtube_dl.utils import (
encode_base_n, encode_base_n,
caesar, caesar,
clean_html, clean_html,
clean_podcast_url,
date_from_str, date_from_str,
DateRange, DateRange,
detect_exe_version, detect_exe_version,
@ -1470,6 +1471,10 @@ Line 1
self.assertEqual(get_elements_by_attribute('class', 'foo', html), []) self.assertEqual(get_elements_by_attribute('class', 'foo', html), [])
self.assertEqual(get_elements_by_attribute('class', 'no-such-foo', html), []) self.assertEqual(get_elements_by_attribute('class', 'no-such-foo', html), [])
def test_clean_podcast_url(self):
self.assertEqual(clean_podcast_url('https://www.podtrac.com/pts/redirect.mp3/chtbl.com/track/5899E/traffic.megaphone.fm/HSW7835899191.mp3'), 'https://traffic.megaphone.fm/HSW7835899191.mp3')
self.assertEqual(clean_podcast_url('https://play.podtrac.com/npr-344098539/edge1.pod.npr.org/anon.npr-podcasts/podcast/npr/waitwait/2020/10/20201003_waitwait_wwdtmpodcast201003-015621a5-f035-4eca-a9a1-7c118d90bc3c.mp3'), 'https://edge1.pod.npr.org/anon.npr-podcasts/podcast/npr/waitwait/2020/10/20201003_waitwait_wwdtmpodcast201003-015621a5-f035-4eca-a9a1-7c118d90bc3c.mp3')
if __name__ == '__main__': if __name__ == '__main__':
unittest.main() unittest.main()

View file

@ -5706,3 +5706,20 @@ def random_birthday(year_field, month_field, day_field):
month_field: str(random_date.month), month_field: str(random_date.month),
day_field: str(random_date.day), day_field: str(random_date.day),
} }
def clean_podcast_url(url):
return re.sub(r'''(?x)
(?:
(?:
chtbl\.com/track|
media\.blubrry\.com| # https://create.blubrry.com/resources/podcast-media-download-statistics/getting-started/
play\.podtrac\.com
)/[^/]+|
(?:dts|www)\.podtrac\.com/(?:pts/)?redirect\.[0-9a-z]{3,4}| # http://analytics.podtrac.com/how-to-measure
flex\.acast\.com|
pd(?:
cn\.co| # https://podcorn.com/analytics-prefix/
st\.fm # https://podsights.com/docs/
)/e
)/''', '', url)