github-com-ytdl-org/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-01-07 17:16:08 +00:00
Code Issues Releases Wiki Activity

[liveleak] Fix extraction (#12053)

Browse source
This commit is contained in:
Yen Chi Hsuan 2017-05-12 19:15:33 +08:00
parent 7274f3d0e9
commit 116283ff64
No known key found for this signature in database
GPG key ID: 7F902A182457CA23
2 changed files with 34 additions and 50 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
ChangeLog
View file

@ -1,6 +1,7 @@
version <unreleased> version <unreleased>
Extractor Extractor
* [liveleak] Fix extraction (#12053)
+ [douyu] Support Douyu shows (#12228) + [douyu] Support Douyu shows (#12228)

83
youtube_dl/extractor/liveleak.py
View file

@ -1,6 +1,5 @@
from __future__ import unicode_literals from __future__ import unicode_literals
import json
import re import re
from .common import InfoExtractor from .common import InfoExtractor
@ -11,10 +10,10 @@ class LiveLeakIE(InfoExtractor):
_VALID_URL = r'https?://(?:\w+\.)?liveleak\.com/view\?(?:.*?)i=(?P<id>[\w_]+)(?:.*)' _VALID_URL = r'https?://(?:\w+\.)?liveleak\.com/view\?(?:.*?)i=(?P<id>[\w_]+)(?:.*)'
_TESTS = [{ _TESTS = [{
'url': 'http://www.liveleak.com/view?i=757_1364311680', 'url': 'http://www.liveleak.com/view?i=757_1364311680',
'md5': '50f79e05ba149149c1b4ea961223d5b3', 'md5': '0813c2430bea7a46bf13acf3406992f4',
'info_dict': { 'info_dict': {
'id': '757_1364311680', 'id': '757_1364311680',
'ext': 'flv', 'ext': 'mp4',
'description': 'extremely bad day for this guy..!', 'description': 'extremely bad day for this guy..!',
'uploader': 'ljfriel2', 'uploader': 'ljfriel2',
'title': 'Most unlucky car accident', 'title': 'Most unlucky car accident',
@ -22,7 +21,7 @@ class LiveLeakIE(InfoExtractor):
} }
}, { }, {
'url': 'http://www.liveleak.com/view?i=f93_1390833151', 'url': 'http://www.liveleak.com/view?i=f93_1390833151',
'md5': 'b13a29626183c9d33944e6a04f41aafc', 'md5': 'd3f1367d14cc3c15bf24fbfbe04b9abf',
'info_dict': { 'info_dict': {
'id': 'f93_1390833151', 'id': 'f93_1390833151',
'ext': 'mp4', 'ext': 'mp4',
@ -32,6 +31,7 @@ class LiveLeakIE(InfoExtractor):
'thumbnail': r're:^https?://.*\.jpg$' 'thumbnail': r're:^https?://.*\.jpg$'
} }
}, { }, {
# Prochan embed
'url': 'http://www.liveleak.com/view?i=4f7_1392687779', 'url': 'http://www.liveleak.com/view?i=4f7_1392687779',
'md5': '42c6d97d54f1db107958760788c5f48f', 'md5': '42c6d97d54f1db107958760788c5f48f',
'info_dict': { 'info_dict': {
@ -41,11 +41,13 @@ class LiveLeakIE(InfoExtractor):
'uploader': 'CapObveus', 'uploader': 'CapObveus',
'title': 'Man is Fatally Struck by Reckless Car While Packing up a Moving Truck', 'title': 'Man is Fatally Struck by Reckless Car While Packing up a Moving Truck',
'age_limit': 18, 'age_limit': 18,
} },
'skip': 'Video is dead',
}, { }, {
# Covers https://github.com/rg3/youtube-dl/pull/5983 # Covers https://github.com/rg3/youtube-dl/pull/5983
# Multiple resolutions
'url': 'http://www.liveleak.com/view?i=801_1409392012', 'url': 'http://www.liveleak.com/view?i=801_1409392012',
'md5': '0b3bec2d888c20728ca2ad3642f0ef15', 'md5': 'c3a449dbaca5c0d1825caecd52a57d7b',
'info_dict': { 'info_dict': {
'id': '801_1409392012', 'id': '801_1409392012',
'ext': 'mp4', 'ext': 'mp4',
@ -93,57 +95,38 @@ class LiveLeakIE(InfoExtractor):
webpage, 'age limit', default=None)) webpage, 'age limit', default=None))
video_thumbnail = self._og_search_thumbnail(webpage) video_thumbnail = self._og_search_thumbnail(webpage)
sources_raw = self._search_regex( entries = self._parse_html5_media_entries(url, webpage, video_id)
r'(?s)sources:\s*(\[.*?\]),', webpage, 'video URLs', default=None) if not entries:
if sources_raw is None: # Maybe an embed?
alt_source = self._search_regex( embed_url = self._search_regex(
r'(file: ".*?"),', webpage, 'video URL', default=None) r'<iframe[^>]+src="((?:https?:)?//(?:www\.)?(?:prochan|youtube)\.com/embed[^"]+)"',
if alt_source: webpage, 'embed URL')
sources_raw = '[{ %s}]' % alt_source return {
else: '_type': 'url_transparent',
# Maybe an embed? 'url': embed_url,
embed_url = self._search_regex( 'id': video_id,
r'<iframe[^>]+src="(https?://(?:www\.)?(?:prochan|youtube)\.com/embed[^"]+)"', 'title': video_title,
webpage, 'embed URL') 'description': video_description,
return { 'uploader': video_uploader,
'_type': 'url_transparent', 'age_limit': age_limit,
'url': embed_url, }
'id': video_id,
'title': video_title,
'description': video_description,
'uploader': video_uploader,
'age_limit': age_limit,
}
sources_json = re.sub(r'\s([a-z]+):\s', r'"\1": ', sources_raw) info_dict = entries[0]
sources = json.loads(sources_json)
formats = [{ for a_format in info_dict['formats']:
'format_id': '%s' % i, if not a_format.get('height'):
'format_note': s.get('label'), a_format['height'] = self._search_regex(
'url': s['file'], r'([0-9]+)p\.mp4', a_format['url'], 'height label', default=None)
} for i, s in enumerate(sources)]
for i, s in enumerate(sources): self._sort_formats(info_dict['formats'])
# Removing '.h264_*.mp4' gives the raw video, which is essentially
# the same video without the LiveLeak logo at the top (see
# https://github.com/rg3/youtube-dl/pull/4768)
orig_url = re.sub(r'\.h264_.+?\.mp4', '', s['file'])
if s['file'] != orig_url:
formats.append({
'format_id': 'original-%s' % i,
'format_note': s.get('label'),
'url': orig_url,
'preference': 1,
})
self._sort_formats(formats)
return { info_dict.update({
'id': video_id, 'id': video_id,
'title': video_title, 'title': video_title,
'description': video_description, 'description': video_description,
'uploader': video_uploader, 'uploader': video_uploader,
'formats': formats,
'age_limit': age_limit, 'age_limit': age_limit,
'thumbnail': video_thumbnail, 'thumbnail': video_thumbnail,
} })
return info_dict