mirror of
https://codeberg.org/polarisfm/youtube-dl
synced 2024-11-26 10:24:33 +01:00
[generic] Add support for RSS feeds (Fixes #667)
This commit is contained in:
parent
280bc5dad6
commit
4fc946b546
@ -250,5 +250,14 @@ class TestPlaylists(unittest.TestCase):
|
||||
self.assertEqual(result['title'], 'python language')
|
||||
self.assertTrue(len(result['entries']) == 15)
|
||||
|
||||
def test_generic_rss_feed(self):
|
||||
dl = FakeYDL()
|
||||
ie = GenericIE(dl)
|
||||
result = ie.extract('http://www.escapistmagazine.com/rss/videos/list/1.xml')
|
||||
self.assertIsPlaylist(result)
|
||||
self.assertEqual(result['id'], 'http://www.escapistmagazine.com/rss/videos/list/1.xml')
|
||||
self.assertEqual(result['title'], 'Zero Punctuation')
|
||||
self.assertTrue(len(result['entries']) > 10)
|
||||
|
||||
if __name__ == '__main__':
|
||||
unittest.main()
|
||||
|
@ -4,6 +4,7 @@ from __future__ import unicode_literals
|
||||
|
||||
import os
|
||||
import re
|
||||
import xml.etree.ElementTree
|
||||
|
||||
from .common import InfoExtractor
|
||||
from .youtube import YoutubeIE
|
||||
@ -159,6 +160,25 @@ class GenericIE(InfoExtractor):
|
||||
raise ExtractorError('Invalid URL protocol')
|
||||
return response
|
||||
|
||||
def _extract_rss(self, url, video_id, doc):
|
||||
playlist_title = doc.find('./channel/title').text
|
||||
playlist_desc_el = doc.find('./channel/description')
|
||||
playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
|
||||
|
||||
entries = [{
|
||||
'_type': 'url',
|
||||
'url': e.find('link').text,
|
||||
'title': e.find('title').text,
|
||||
} for e in doc.findall('./channel/item')]
|
||||
|
||||
return {
|
||||
'_type': 'playlist',
|
||||
'id': url,
|
||||
'title': playlist_title,
|
||||
'description': playlist_desc,
|
||||
'entries': entries,
|
||||
}
|
||||
|
||||
def _real_extract(self, url):
|
||||
parsed_url = compat_urlparse.urlparse(url)
|
||||
if not parsed_url.scheme:
|
||||
@ -219,6 +239,14 @@ class GenericIE(InfoExtractor):
|
||||
|
||||
self.report_extraction(video_id)
|
||||
|
||||
# Is it an RSS feed?
|
||||
try:
|
||||
doc = xml.etree.ElementTree.fromstring(webpage)
|
||||
if doc.tag == 'rss':
|
||||
return self._extract_rss(url, video_id, doc)
|
||||
except xml.etree.ElementTree.ParseError:
|
||||
pass
|
||||
|
||||
# it's tempting to parse this further, but you would
|
||||
# have to take into account all the variations like
|
||||
# Video Title - Site Name
|
||||
|
Loading…
Reference in New Issue
Block a user