MarcoBuster/youtube-dl
1
1
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-11-27 21:46:54 +01:00
Code Issues Releases Wiki Activity

[dhm] Improve _VALID_URL and add test

Browse Source
This commit is contained in:
Sergey M․ 2015-03-28 23:55:15 +06:00
parent b7a2268e7b
commit 5a3b315b5f
1 changed files with 15 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
youtube_dl/extractor/dhm.py
View File

@ -9,9 +9,9 @@ from ..utils import (
class DHMIE(InfoExtractor): class DHMIE(InfoExtractor):
IE_DESC = 'Filmarchiv - Deutsches Historisches Museum' 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/', 'url': 'http://www.dhm.de/filmarchiv/die-filme/the-marshallplan-at-work-in-west-germany/',
'md5': '11c475f670209bf6acca0b2b7ef51827', 'md5': '11c475f670209bf6acca0b2b7ef51827',
'info_dict': { 'info_dict': {
@ -21,8 +21,17 @@ class DHMIE(InfoExtractor):
'description': 'md5:1fabd480c153f97b07add61c44407c82', 'description': 'md5:1fabd480c153f97b07add61c44407c82',
'duration': 660, 'duration': 660,
'thumbnail': 're:^https?://.*\.jpg$', 'thumbnail': 're:^https?://.*\.jpg$',
} },
} }, {
'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): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)
@ -49,10 +58,10 @@ class DHMIE(InfoExtractor):
webpage, 'title').strip() webpage, 'title').strip()
description = self._html_search_regex( description = self._html_search_regex(
r'<p><strong>Description:</strong>(.+?)</p>', r'<p><strong>Description:</strong>(.+?)</p>',
webpage, 'description', fatal=False) webpage, 'description', default=None)
duration = parse_duration(self._search_regex( duration = parse_duration(self._search_regex(
r'<em>Length\s*</em>\s*:\s*</strong>([^<]+)', r'<em>Length\s*</em>\s*:\s*</strong>([^<]+)',
webpage, 'duration', fatal=False)) webpage, 'duration', default=None))
return { return {
'id': video_id, 'id': video_id,