1
0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-28 09:30:57 +01:00

[xuite] Move to the new HTML5 API and reduce # of requests

This commit is contained in:
Yen Chi Hsuan 2017-06-13 16:21:26 +08:00
parent 609ff8ca19
commit babbc04d45

View File

@ -1,14 +1,13 @@
# coding: utf-8
from __future__ import unicode_literals
import base64
from .common import InfoExtractor
from ..compat import compat_urllib_parse_unquote
from ..utils import (
ExtractorError,
float_or_none,
get_element_by_attribute,
parse_iso8601,
parse_duration,
remove_end,
)
@ -24,6 +23,7 @@ class XuiteIE(InfoExtractor):
'id': '3860914',
'ext': 'mp3',
'title': '孤單南半球-歐德陽',
'description': '孤單南半球-歐德陽',
'thumbnail': r're:^https?://.*\.jpg$',
'duration': 247.246,
'timestamp': 1314932940,
@ -44,7 +44,7 @@ class XuiteIE(InfoExtractor):
'duration': 596.458,
'timestamp': 1454242500,
'upload_date': '20160131',
'uploader': 'yan12125',
'uploader': '屁姥',
'uploader_id': '12158353',
'categories': ['個人短片'],
'description': 'http://download.blender.org/peach/bigbuckbunny_movies/BigBuckBunny_320x180.mp4',
@ -72,10 +72,10 @@ class XuiteIE(InfoExtractor):
# from http://forgetfulbc.blogspot.com/2016/06/date.html
'url': 'http://vlog.xuite.net/embed/cE1xbENoLTI3NDQ3MzM2LmZsdg==?ar=0&as=0',
'info_dict': {
'id': 'cE1xbENoLTI3NDQ3MzM2LmZsdg==',
'id': '27447336',
'ext': 'mp4',
'title': '男女平權只是口號?專家解釋約會時男生是否該幫女生付錢 (中字)',
'description': 'md5:f0abdcb69df300f522a5442ef3146f2a',
'description': 'md5:1223810fa123b179083a3aed53574706',
'timestamp': 1466160960,
'upload_date': '20160617',
'uploader': 'B.C. & Lowy',
@ -86,29 +86,9 @@ class XuiteIE(InfoExtractor):
'only_matching': True,
}]
@staticmethod
def base64_decode_utf8(data):
return base64.b64decode(data.encode('utf-8')).decode('utf-8')
@staticmethod
def base64_encode_utf8(data):
return base64.b64encode(data.encode('utf-8')).decode('utf-8')
def _extract_flv_config(self, encoded_media_id):
flv_config = self._download_xml(
'http://vlog.xuite.net/flash/player?media=%s' % encoded_media_id,
'flv config')
prop_dict = {}
for prop in flv_config.findall('./property'):
prop_id = self.base64_decode_utf8(prop.attrib['id'])
# CDATA may be empty in flv config
if not prop.text:
continue
encoded_content = self.base64_decode_utf8(prop.text)
prop_dict[prop_id] = compat_urllib_parse_unquote(encoded_content)
return prop_dict
def _real_extract(self, url):
# /play/ URLs provide embedded video URL and more metadata
url = url.replace('/embed/', '/play/')
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
@ -121,51 +101,53 @@ class XuiteIE(InfoExtractor):
'%s returned error: %s' % (self.IE_NAME, error_msg),
expected=True)
encoded_media_id = self._search_regex(
r'attributes\.name\s*=\s*"([^"]+)"', webpage,
'encoded media id', default=None)
if encoded_media_id is None:
video_id = self._html_search_regex(
r'data-mediaid="(\d+)"', webpage, 'media id')
encoded_media_id = self.base64_encode_utf8(video_id)
flv_config = self._extract_flv_config(encoded_media_id)
media_info = self._parse_json(self._search_regex(
r'var\s+mediaInfo\s*=\s*({.*});', webpage, 'media info'), video_id)
FORMATS = {
'audio': 'mp3',
'video': 'mp4',
}
video_id = media_info['MEDIA_ID']
formats = []
for format_tag in ('src', 'hq_src'):
video_url = flv_config.get(format_tag)
for key in ('html5Url', 'html5HQUrl'):
video_url = media_info.get(key)
if not video_url:
continue
format_id = self._search_regex(
r'\bq=(.+?)\b', video_url, 'format id', default=format_tag)
r'\bq=(.+?)\b', video_url, 'format id', default=None)
formats.append({
'url': video_url,
'ext': FORMATS.get(flv_config['type'], 'mp4'),
'ext': 'mp4' if format_id.isnumeric() else format_id,
'format_id': format_id,
'height': int(format_id) if format_id.isnumeric() else None,
})
self._sort_formats(formats)
timestamp = flv_config.get('publish_datetime')
timestamp = media_info.get('PUBLISH_DATETIME')
if timestamp:
timestamp = parse_iso8601(timestamp + ' +0800', ' ')
category = flv_config.get('category')
category = media_info.get('catName')
categories = [category] if category else []
uploader = media_info.get('NICKNAME')
uploader_url = None
author_div = get_element_by_attribute('itemprop', 'author', webpage)
if author_div:
uploader = uploader or self._html_search_meta('name', author_div)
uploader_url = self._html_search_regex(
r'<link[^>]+itemprop="url"[^>]+href="([^"]+)"', author_div,
'uploader URL', fatal=False)
return {
'id': video_id,
'title': flv_config['title'],
'description': flv_config.get('description'),
'thumbnail': flv_config.get('thumb'),
'title': media_info['TITLE'],
'description': remove_end(media_info.get('metaDesc'), ' (Xuite 影音)'),
'thumbnail': media_info.get('ogImageUrl'),
'timestamp': timestamp,
'uploader': flv_config.get('author_name'),
'uploader_id': flv_config.get('author_id'),
'duration': parse_duration(flv_config.get('duration')),
'uploader': uploader,
'uploader_id': media_info.get('MEMBER_ID'),
'uploader_url': uploader_url,
'duration': float_or_none(media_info.get('MEDIA_DURATION'), 1000000),
'categories': categories,
'formats': formats,
}