mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-11-22 02:15:12 +00:00
86e5f3ed2e
Using https://github.com/asottile/pyupgrade 1. `__future__` imports and `coding: utf-8` were removed 2. Files were rewritten with `pyupgrade --py36-plus --keep-percent-format` 3. f-strings were cherry-picked from `pyupgrade --py36-plus` Extractors are left untouched (except removing header) to avoid unnecessary merge conflicts
75 lines
2.4 KiB
Python
75 lines
2.4 KiB
Python
import re
|
|
|
|
from .common import InfoExtractor
|
|
from ..utils import (
|
|
ExtractorError,
|
|
urlencode_postdata,
|
|
)
|
|
|
|
|
|
class StreamcloudIE(InfoExtractor):
|
|
IE_NAME = 'streamcloud.eu'
|
|
_VALID_URL = r'https?://streamcloud\.eu/(?P<id>[a-zA-Z0-9_-]+)(?:/(?P<fname>[^#?]*)\.html)?'
|
|
|
|
_TESTS = [{
|
|
'url': 'http://streamcloud.eu/skp9j99s4bpz/youtube-dl_test_video_____________-BaW_jenozKc.mp4.html',
|
|
'md5': '6bea4c7fa5daaacc2a946b7146286686',
|
|
'info_dict': {
|
|
'id': 'skp9j99s4bpz',
|
|
'ext': 'mp4',
|
|
'title': 'youtube-dl test video \'/\\ ä ↭',
|
|
},
|
|
'skip': 'Only available from the EU'
|
|
}, {
|
|
'url': 'http://streamcloud.eu/ua8cmfh1nbe6/NSHIP-148--KUC-NG--H264-.mp4.html',
|
|
'only_matching': True,
|
|
}]
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
url = 'http://streamcloud.eu/%s' % video_id
|
|
|
|
orig_webpage = self._download_webpage(url, video_id)
|
|
|
|
if '>File Not Found<' in orig_webpage:
|
|
raise ExtractorError(
|
|
'Video %s does not exist' % video_id, expected=True)
|
|
|
|
fields = re.findall(r'''(?x)<input\s+
|
|
type="(?:hidden|submit)"\s+
|
|
name="([^"]+)"\s+
|
|
(?:id="[^"]+"\s+)?
|
|
value="([^"]*)"
|
|
''', orig_webpage)
|
|
|
|
self._sleep(6, video_id)
|
|
|
|
webpage = self._download_webpage(
|
|
url, video_id, data=urlencode_postdata(fields), headers={
|
|
b'Content-Type': b'application/x-www-form-urlencoded',
|
|
})
|
|
|
|
try:
|
|
title = self._html_search_regex(
|
|
r'<h1[^>]*>([^<]+)<', webpage, 'title')
|
|
video_url = self._search_regex(
|
|
r'file:\s*"([^"]+)"', webpage, 'video URL')
|
|
except ExtractorError:
|
|
message = self._html_search_regex(
|
|
r'(?s)<div[^>]+class=(["\']).*?msgboxinfo.*?\1[^>]*>(?P<message>.+?)</div>',
|
|
webpage, 'message', default=None, group='message')
|
|
if message:
|
|
raise ExtractorError('%s said: %s' % (self.IE_NAME, message), expected=True)
|
|
raise
|
|
thumbnail = self._search_regex(
|
|
r'image:\s*"([^"]+)"', webpage, 'thumbnail URL', fatal=False)
|
|
|
|
return {
|
|
'id': video_id,
|
|
'title': title,
|
|
'url': video_url,
|
|
'thumbnail': thumbnail,
|
|
'http_headers': {
|
|
'Referer': url,
|
|
},
|
|
}
|