X-Git-Url: http://git.bitcoin.ninja/index.cgi?p=youtube-dl;a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fsohu.py;h=30760ca06be4b3fc112f3fe0200c74b665d64855;hp=ea8fc258d1e7f361bc2a75360dc4c65774674622;hb=dcdb292fddc82ae11f4c0b647815a45c88a6b6d5;hpb=974c1b2d4292308a26a47136e7fcf9b61f8b285a diff --git a/youtube_dl/extractor/sohu.py b/youtube_dl/extractor/sohu.py index ea8fc258d..30760ca06 100644 --- a/youtube_dl/extractor/sohu.py +++ b/youtube_dl/extractor/sohu.py @@ -1,4 +1,4 @@ -# encoding: utf-8 +# coding: utf-8 from __future__ import unicode_literals import re @@ -6,21 +6,18 @@ import re from .common import InfoExtractor from ..compat import ( compat_str, - compat_urllib_parse, -) -from ..utils import ( - ExtractorError, - sanitized_Request, + compat_urllib_parse_urlencode, ) +from ..utils import ExtractorError class SohuIE(InfoExtractor): _VALID_URL = r'https?://(?Pmy\.)?tv\.sohu\.com/.+?/(?(mytv)|n)(?P\d+)\.shtml.*?' + # Sohu videos give different MD5 sums on Travis CI and my machine _TESTS = [{ 'note': 'This video is available only in Mainland China', 'url': 'http://tv.sohu.com/20130724/n382479172.shtml#super', - 'md5': '29175c8cadd8b5cc4055001e85d6b372', 'info_dict': { 'id': '382479172', 'ext': 'mp4', @@ -29,7 +26,6 @@ class SohuIE(InfoExtractor): 'skip': 'On available in China', }, { 'url': 'http://tv.sohu.com/20150305/n409385080.shtml', - 'md5': '699060e75cf58858dd47fb9c03c42cfb', 'info_dict': { 'id': '409385080', 'ext': 'mp4', @@ -37,7 +33,6 @@ class SohuIE(InfoExtractor): } }, { 'url': 'http://my.tv.sohu.com/us/232799889/78693464.shtml', - 'md5': '9bf34be48f2f4dadcb226c74127e203c', 'info_dict': { 'id': '78693464', 'ext': 'mp4', @@ -51,7 +46,6 @@ class SohuIE(InfoExtractor): 'title': '【神探苍实战秘籍】第13期 战争之影 赫卡里姆', }, 'playlist': [{ - 'md5': 'bdbfb8f39924725e6589c146bc1883ad', 'info_dict': { 'id': '78910339_part1', 'ext': 'mp4', @@ -59,7 +53,6 @@ class SohuIE(InfoExtractor): 'title': '【神探苍实战秘籍】第13期 战争之影 赫卡里姆', } }, { - 'md5': '3e1f46aaeb95354fd10e7fca9fc1804e', 'info_dict': { 'id': '78910339_part2', 'ext': 'mp4', @@ -67,7 +60,6 @@ class SohuIE(InfoExtractor): 'title': '【神探苍实战秘籍】第13期 战争之影 赫卡里姆', } }, { - 'md5': '8407e634175fdac706766481b9443450', 'info_dict': { 'id': '78910339_part3', 'ext': 'mp4', @@ -96,15 +88,10 @@ class SohuIE(InfoExtractor): else: base_data_url = 'http://hot.vrs.sohu.com/vrs_flash.action?vid=' - req = sanitized_Request(base_data_url + vid_id) - - cn_verification_proxy = self._downloader.params.get('cn_verification_proxy') - if cn_verification_proxy: - req.add_header('Ytdl-request-proxy', cn_verification_proxy) - return self._download_json( - req, video_id, - 'Downloading JSON data for %s' % vid_id) + base_data_url + vid_id, video_id, + 'Downloading JSON data for %s' % vid_id, + headers=self.geo_verification_headers()) mobj = re.match(self._VALID_URL, url) video_id = mobj.group('id') @@ -170,7 +157,7 @@ class SohuIE(InfoExtractor): if retries > 0: download_note += ' (retry #%d)' % retries part_info = self._parse_json(self._download_webpage( - 'http://%s/?%s' % (allot, compat_urllib_parse.urlencode(params)), + 'http://%s/?%s' % (allot, compat_urllib_parse_urlencode(params)), video_id, download_note), video_id) video_url = part_info['url']