midgard/youtube-dl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[escapist] Add support for HD format (Closes #1755)

Browse source
This commit is contained in:
Philipp Hagemeister 2013-11-20 06:47:50 +01:00
parent 0a120f74b2
commit 100959a6d9
1 changed files with 34 additions and 23 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

57
youtube_dl/extractor/escapist.py
View file

@ -11,11 +11,11 @@ from ..utils import (
class EscapistIE(InfoExtractor): class EscapistIE(InfoExtractor):
_VALID_URL = r'^(https?://)?(www\.)?escapistmagazine\.com/videos/view/(?P<showname>[^/]+)/(?P<episode>[^/?]+)[/?]?.*$' _VALID_URL = r'^https?://?(www\.)?escapistmagazine\.com/videos/view/(?P<showname>[^/]+)/(?P<episode>[^/?]+)[/?]?.*$'
_TEST = { _TEST = {
u'url': u'http://www.escapistmagazine.com/videos/view/the-escapist-presents/6618-Breaking-Down-Baldurs-Gate', u'url': u'http://www.escapistmagazine.com/videos/view/the-escapist-presents/6618-Breaking-Down-Baldurs-Gate',
u'file': u'6618-Breaking-Down-Baldurs-Gate.mp4', u'file': u'6618-Breaking-Down-Baldurs-Gate.mp4',
u'md5': u'c6793dbda81388f4264c1ba18684a74d', u'md5': u'ab3a706c681efca53f0a35f1415cf0d1',
u'info_dict': { u'info_dict': {
u"description": u"Baldur's Gate: Original, Modded or Enhanced Edition? I'll break down what you can expect from the new Baldur's Gate: Enhanced Edition.", u"description": u"Baldur's Gate: Original, Modded or Enhanced Edition? I'll break down what you can expect from the new Baldur's Gate: Enhanced Edition.",
u"uploader": u"the-escapist-presents", u"uploader": u"the-escapist-presents",
@ -25,50 +25,61 @@ class EscapistIE(InfoExtractor):
def _real_extract(self, url): def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url) mobj = re.match(self._VALID_URL, url)
if mobj is None:
raise ExtractorError(u'Invalid URL: %s' % url)
showName = mobj.group('showname') showName = mobj.group('showname')
videoId = mobj.group('episode') videoId = mobj.group('episode')
self.report_extraction(videoId) self.report_extraction(videoId)
webpage = self._download_webpage(url, videoId) webpage = self._download_webpage(url, videoId)
videoDesc = self._html_search_regex('<meta name="description" content="([^"]*)"', videoDesc = self._html_search_regex(
r'<meta name="description" content="([^"]*)"',
webpage, u'description', fatal=False) webpage, u'description', fatal=False)
playerUrl = self._og_search_video_url(webpage, name='player url') playerUrl = self._og_search_video_url(webpage, name='player url')
title = self._html_search_regex('<meta name="title" content="([^"]*)"', title = self._html_search_regex(
'r<meta name="title" content="([^"]*)"',
webpage, u'player url').split(' : ')[-1] webpage, u'player url').split(' : ')[-1]
configUrl = self._search_regex('config=(.*)$', playerUrl, u'config url') configUrl = self._search_regex('config=(.*)$', playerUrl, u'config url')
configUrl = compat_urllib_parse.unquote(configUrl) configUrl = compat_urllib_parse.unquote(configUrl)
configJSON = self._download_webpage(configUrl, videoId, formats = []
u'Downloading configuration',
u'unable to download configuration')
# Technically, it's JavaScript, not JSON def _add_format(name, cfgurl):
configJSON = configJSON.replace("'", '"') configJSON = self._download_webpage(
cfgurl, videoId,
u'Downloading ' + name + ' configuration',
u'Unable to download ' + name + ' configuration')
# Technically, it's JavaScript, not JSON
configJSON = configJSON.replace("'", '"')
try:
config = json.loads(configJSON)
except (ValueError,) as err:
raise ExtractorError(u'Invalid JSON in configuration file: ' + compat_str(err))
playlist = config['playlist']
formats.append({
'url': playlist[1]['url'],
'format_id': name,
})
_add_format(u'normal', configUrl)
try: try:
config = json.loads(configJSON) hq_url = (configUrl +
except (ValueError,) as err: ('&hq=1' if '?' in configUrl else configUrl + '?hq=1'))
raise ExtractorError(u'Invalid JSON in configuration file: ' + compat_str(err)) try:
_add_format(u'hq', hq_url)
except ExtractorError:
pass # That's fine, we'll just use normal quality
playlist = config['playlist'] return {
videoUrl = playlist[1]['url']
info = {
'id': videoId, 'id': videoId,
'url': videoUrl, 'formats': formats,
'uploader': showName, 'uploader': showName,
'upload_date': None,
'title': title, 'title': title,
'ext': 'mp4',
'thumbnail': self._og_search_thumbnail(webpage), 'thumbnail': self._og_search_thumbnail(webpage),
'description': videoDesc, 'description': videoDesc,
'player_url': playerUrl, 'player_url': playerUrl,
} }
return [info]