[ro220] Fix broken extractor and modernize (#4054)

This commit is contained in:
Naglis Jonaitis 2014-10-30 01:42:52 +02:00
parent feb74960eb
commit b8a618f898
1 changed files with 23 additions and 23 deletions

View File

@ -1,43 +1,43 @@
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
clean_html,
compat_parse_qs,
)
from ..utils import compat_urllib_parse_unquote
class Ro220IE(InfoExtractor):
IE_NAME = '220.ro'
_VALID_URL = r'(?x)(?:https?://)?(?:www\.)?220\.ro/(?P<category>[^/]+)/(?P<shorttitle>[^/]+)/(?P<video_id>[^/]+)'
_VALID_URL = r'(?x)(?:https?://)?(?:www\.)?220\.ro/(?P<category>[^/]+)/(?P<shorttitle>[^/]+)/(?P<id>[^/]+)'
_TEST = {
"url": "http://www.220.ro/sport/Luati-Le-Banii-Sez-4-Ep-1/LYV6doKo7f/",
'file': 'LYV6doKo7f.mp4',
'url': 'http://www.220.ro/sport/Luati-Le-Banii-Sez-4-Ep-1/LYV6doKo7f/',
'md5': '03af18b73a07b4088753930db7a34add',
'info_dict': {
"title": "Luati-le Banii sez 4 ep 1",
"description": "re:^Iata-ne reveniti dupa o binemeritata vacanta\. +Va astept si pe Facebook cu pareri si comentarii.$",
'id': 'LYV6doKo7f',
'ext': 'mp4',
'title': 'Luati-le Banii sez 4 ep 1',
'description': 're:^Iata-ne reveniti dupa o binemeritata vacanta\. +Va astept si pe Facebook cu pareri si comentarii.$',
}
}
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('video_id')
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
flashVars_str = self._search_regex(
r'<param name="flashVars" value="([^"]+)"',
webpage, 'flashVars')
flashVars = compat_parse_qs(flashVars_str)
url = compat_urllib_parse_unquote(self._search_regex(
r'(?s)clip\s*:\s*{.*?url\s*:\s*\'([^\']+)\'', webpage, 'url'))
title = self._og_search_title(webpage)
description = self._og_search_description(webpage)
thumbnail = self._og_search_thumbnail(webpage)
formats = [{
'format_id': 'sd',
'url': url,
'ext': 'mp4',
}]
return {
'_type': 'video',
'id': video_id,
'ext': 'mp4',
'url': flashVars['videoURL'][0],
'title': flashVars['title'][0],
'description': clean_html(flashVars['desc'][0]),
'thumbnail': flashVars['preview'][0],
'formats': formats,
'title': title,
'description': description,
'thumbnail': thumbnail,
}