Merge branch 'naglis-jove'

This commit is contained in:
Sergey M․ 2014-08-19 20:02:31 +07:00
commit 696d49815e
3 changed files with 82 additions and 0 deletions

View File

@ -147,6 +147,7 @@ from .ivi import (
from .izlesene import IzleseneIE from .izlesene import IzleseneIE
from .jadorecettepub import JadoreCettePubIE from .jadorecettepub import JadoreCettePubIE
from .jeuxvideo import JeuxVideoIE from .jeuxvideo import JeuxVideoIE
from .jove import JoveIE
from .jukebox import JukeboxIE from .jukebox import JukeboxIE
from .justintv import JustinTVIE from .justintv import JustinTVIE
from .jpopsukitv import JpopsukiIE from .jpopsukitv import JpopsukiIE

View File

@ -0,0 +1,80 @@
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
unified_strdate
)
class JoveIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?jove\.com/video/(?P<id>[0-9]+)'
_CHAPTERS_URL = 'http://www.jove.com/video-chapters?videoid={video_id:}'
_TESTS = [
{
'url': 'http://www.jove.com/video/2744/electrode-positioning-montage-transcranial-direct-current',
'md5': '93723888d82dbd6ba8b3d7d0cd65dd2b',
'info_dict': {
'id': '2744',
'ext': 'mp4',
'title': 'Electrode Positioning and Montage in Transcranial Direct Current Stimulation',
'description': 'md5:015dd4509649c0908bc27f049e0262c6',
'thumbnail': 're:^https?://.*\.png$',
'upload_date': '20110523',
}
},
{
'url': 'http://www.jove.com/video/51796/culturing-caenorhabditis-elegans-axenic-liquid-media-creation',
'md5': '914aeb356f416811d911996434811beb',
'info_dict': {
'id': '51796',
'ext': 'mp4',
'title': 'Culturing Caenorhabditis elegans in Axenic Liquid Media and Creation of Transgenic Worms by Microparticle Bombardment',
'description': 'md5:35ff029261900583970c4023b70f1dc9',
'thumbnail': 're:^https?://.*\.png$',
'upload_date': '20140802',
}
},
]
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
webpage = self._download_webpage(url, video_id)
chapters_id = self._html_search_regex(
r'/video-chapters\?videoid=([0-9]+)', webpage, 'chapters id')
chapters_xml = self._download_xml(
self._CHAPTERS_URL.format(video_id=chapters_id),
video_id, note='Downloading chapters XML',
errnote='Failed to download chapters XML')
video_url = chapters_xml.attrib.get('video')
if not video_url:
raise ExtractorError('Failed to get the video URL')
title = self._html_search_meta('citation_title', webpage, 'title')
thumbnail = self._og_search_thumbnail(webpage)
description = self._html_search_regex(
r'<div id="section_body_summary"><p class="jove_content">(.+?)</p>',
webpage, 'description', fatal=False)
publish_date = unified_strdate(self._html_search_meta(
'citation_publication_date', webpage, 'publish date', fatal=False))
comment_count = self._html_search_regex(
r'<meta name="num_comments" content="(\d+) Comments?"',
webpage, 'comment count', fatal=False)
return {
'id': video_id,
'title': title,
'url': video_url,
'thumbnail': thumbnail,
'description': description,
'upload_date': publish_date,
'comment_count': comment_count,
}

View File

@ -827,6 +827,7 @@ def unified_strdate(date_str):
'%b %dnd %Y %I:%M%p', '%b %dnd %Y %I:%M%p',
'%b %dth %Y %I:%M%p', '%b %dth %Y %I:%M%p',
'%Y-%m-%d', '%Y-%m-%d',
'%Y/%m/%d',
'%d.%m.%Y', '%d.%m.%Y',
'%d/%m/%Y', '%d/%m/%Y',
'%Y/%m/%d %H:%M:%S', '%Y/%m/%d %H:%M:%S',