aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor/dhm.py
diff options
context:
space:
mode:
authorSergey M․ <dstftw@gmail.com>2015-03-28 23:55:15 +0600
committerSergey M․ <dstftw@gmail.com>2015-03-28 23:55:15 +0600
commit5a3b315b5fcc1b1e153baeaed669e64d137d4047 (patch)
tree939781e61ddc6c9dd0ea757dd757f2f94d46b6b9 /youtube_dl/extractor/dhm.py
parentb7a2268e7b52fbedd1630ad101460d76cca9dcdd (diff)
downloadyoutube-dl-5a3b315b5fcc1b1e153baeaed669e64d137d4047.tar.xz
[dhm] Improve _VALID_URL and add test
Diffstat (limited to 'youtube_dl/extractor/dhm.py')
-rw-r--r--youtube_dl/extractor/dhm.py21
1 files changed, 15 insertions, 6 deletions
diff --git a/youtube_dl/extractor/dhm.py b/youtube_dl/extractor/dhm.py
index 80ee40018..3ed1f1663 100644
--- a/youtube_dl/extractor/dhm.py
+++ b/youtube_dl/extractor/dhm.py
@@ -9,9 +9,9 @@ from ..utils import (
class DHMIE(InfoExtractor):
IE_DESC = 'Filmarchiv - Deutsches Historisches Museum'
- _VALID_URL = r'http://www\.dhm\.de/filmarchiv/die-filme/(?P<id>[^/]+)'
+ _VALID_URL = r'https?://(?:www\.)?dhm\.de/filmarchiv/(?:[^/]+/)+(?P<id>[^/]+)'
- _TEST = {
+ _TESTS = [{
'url': 'http://www.dhm.de/filmarchiv/die-filme/the-marshallplan-at-work-in-west-germany/',
'md5': '11c475f670209bf6acca0b2b7ef51827',
'info_dict': {
@@ -21,8 +21,17 @@ class DHMIE(InfoExtractor):
'description': 'md5:1fabd480c153f97b07add61c44407c82',
'duration': 660,
'thumbnail': 're:^https?://.*\.jpg$',
- }
- }
+ },
+ }, {
+ 'url': 'http://www.dhm.de/filmarchiv/02-mapping-the-wall/peter-g/rolle-1/',
+ 'md5': '09890226332476a3e3f6f2cb74734aa5',
+ 'info_dict': {
+ 'id': 'rolle-1',
+ 'ext': 'flv',
+ 'title': 'ROLLE 1',
+ 'thumbnail': 're:^https?://.*\.jpg$',
+ },
+ }]
def _real_extract(self, url):
video_id = self._match_id(url)
@@ -49,10 +58,10 @@ class DHMIE(InfoExtractor):
webpage, 'title').strip()
description = self._html_search_regex(
r'<p><strong>Description:</strong>(.+?)</p>',
- webpage, 'description', fatal=False)
+ webpage, 'description', default=None)
duration = parse_duration(self._search_regex(
r'<em>Length\s*</em>\s*:\s*</strong>([^<]+)',
- webpage, 'duration', fatal=False))
+ webpage, 'duration', default=None))
return {
'id': video_id,