w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[freesound] Modernize

Browse source
This commit is contained in:
Jaime Marquínez Ferrándiz 2014-02-10 21:03:14 +01:00
parent af6ba6a1c4
commit 34bd987811
1 changed files with 23 additions and 20 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

37
youtube_dl/extractor/freesound.py
View file

@ -1,18 +1,21 @@
from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..utils import determine_ext
class FreesoundIE(InfoExtractor): class FreesoundIE(InfoExtractor):
_VALID_URL = r'(?:https?://)?(?:www\.)?freesound\.org/people/([^/]+)/sounds/(?P<id>[^/]+)' _VALID_URL = r'https?://(?:www\.)?freesound\.org/people/([^/]+)/sounds/(?P<id>[^/]+)'
_TEST = { _TEST = {
u'url': u'http://www.freesound.org/people/miklovan/sounds/194503/', 'url': 'http://www.freesound.org/people/miklovan/sounds/194503/',
u'file': u'194503.mp3', 'md5': '12280ceb42c81f19a515c745eae07650',
u'md5': u'12280ceb42c81f19a515c745eae07650', 'info_dict': {
u'info_dict': { 'id': '194503',
u"title": u"gulls in the city.wav", 'ext': 'mp3',
u"uploader" : u"miklovan", 'title': 'gulls in the city.wav',
u'description': u'the sounds of seagulls in the city', 'uploader': 'miklovan',
'description': 'the sounds of seagulls in the city',
} }
} }
@ -20,17 +23,17 @@ class FreesoundIE(InfoExtractor):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
music_id = mobj.group('id') music_id = mobj.group('id')
webpage = self._download_webpage(url, music_id) webpage = self._download_webpage(url, music_id)
title = self._html_search_regex(r'<div id="single_sample_header">.*?<a href="#">(.+?)</a>', title = self._html_search_regex(
r'<div id="single_sample_header">.*?<a href="#">(.+?)</a>',
webpage, 'music title', flags=re.DOTALL) webpage, 'music title', flags=re.DOTALL)
music_url = self._og_search_property('audio', webpage, 'music url') description = self._html_search_regex(
description = self._html_search_regex(r'<div id="sound_description">(.*?)</div>', r'<div id="sound_description">(.*?)</div>', webpage, 'description',
webpage, 'description', fatal=False, flags=re.DOTALL) fatal=False, flags=re.DOTALL)
return [{ return {
'id': music_id, 'id': music_id,
'title': title, 'title': title,
'url': music_url, 'url': self._og_search_property('audio', webpage, 'music url'),
'uploader': self._og_search_property('audio:artist', webpage, 'music uploader'), 'uploader': self._og_search_property('audio:artist', webpage, 'music uploader'),
'ext': determine_ext(music_url),
'description': description, 'description': description,
}] }