[redtube] Fix formats extraction (closes #14122)

This commit is contained in:
Sergey M․ 2017-09-05 22:45:07 +07:00
parent 6348671c4a
commit 880fa66f4f
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -3,6 +3,7 @@ from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_str
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
int_or_none, int_or_none,
@ -62,7 +63,23 @@ class RedTubeIE(InfoExtractor):
'format_id': format_id, 'format_id': format_id,
'height': int_or_none(format_id), 'height': int_or_none(format_id),
}) })
else: medias = self._parse_json(
self._search_regex(
r'mediaDefinition\s*:\s*(\[.+?\])', webpage,
'media definitions', default='{}'),
video_id, fatal=False)
if medias and isinstance(medias, list):
for media in medias:
format_url = media.get('videoUrl')
if not format_url or not isinstance(format_url, compat_str):
continue
format_id = media.get('quality')
formats.append({
'url': format_url,
'format_id': format_id,
'height': int_or_none(format_id),
})
if not formats:
video_url = self._html_search_regex( video_url = self._html_search_regex(
r'<source src="(.+?)" type="video/mp4">', webpage, 'video URL') r'<source src="(.+?)" type="video/mp4">', webpage, 'video URL')
formats.append({'url': video_url}) formats.append({'url': video_url})
@ -73,7 +90,7 @@ class RedTubeIE(InfoExtractor):
r'<span[^>]+class="added-time"[^>]*>ADDED ([^<]+)<', r'<span[^>]+class="added-time"[^>]*>ADDED ([^<]+)<',
webpage, 'upload date', fatal=False)) webpage, 'upload date', fatal=False))
duration = int_or_none(self._search_regex( duration = int_or_none(self._search_regex(
r'videoDuration\s*:\s*(\d+)', webpage, 'duration', fatal=False)) r'videoDuration\s*:\s*(\d+)', webpage, 'duration', default=None))
view_count = str_to_int(self._search_regex( view_count = str_to_int(self._search_regex(
r'<span[^>]*>VIEWS</span></td>\s*<td>([\d,.]+)', r'<span[^>]*>VIEWS</span></td>\s*<td>([\d,.]+)',
webpage, 'view count', fatal=False)) webpage, 'view count', fatal=False))