mirrors
/
youtube-dl
0
0
Fork 0
Code Releases 1 Activity

Merge pull request #2953 from codesparkle/ndr-regexes-escape-correctly 

[ndr] fix regexes containing illegal characters
main
Sergey M. 2014-05-25 05:42:06 +07:00
parent 9eeb7abc6b 77541837e5
commit 65bab327b4
1 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
youtube_dl/extractor/ndr.py
View File

@ -7,6 +7,7 @@ from .common import InfoExtractor
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
int_or_none, int_or_none,
qualities,
) )
@ -57,7 +58,7 @@ class NDRIE(InfoExtractor):
formats = [] formats = []
mp3_url = re.search(r'''{src:'(?P<audio>[^']+)', type:"audio/mp3"},''', page) mp3_url = re.search(r'''\{src:'(?P<audio>[^']+)', type:"audio/mp3"},''', page)
if mp3_url: if mp3_url:
formats.append({ formats.append({
'url': mp3_url.group('audio'), 'url': mp3_url.group('audio'),
@ -66,15 +67,15 @@ class NDRIE(InfoExtractor):
thumbnail = None thumbnail = None
video_url = re.search(r'''3: {src:'(?P<video>.+?)\.hi\.mp4', type:"video/mp4"},''', page) video_url = re.search(r'''3: \{src:'(?P<video>.+?)\.hi\.mp4', type:"video/mp4"},''', page)
if video_url: if video_url:
thumbnails = re.findall(r'''\d+: {src: "([^"]+)"(?: \|\| '[^']+')?, quality: '([^']+)'}''', page) thumbnails = re.findall(r'''\d+: \{src: "([^"]+)"(?: \|\| '[^']+')?, quality: '([^']+)'}''', page)
if thumbnails: if thumbnails:
QUALITIES = ['xs', 's', 'm', 'l', 'xl'] quality_key = qualities(['xs', 's', 'm', 'l', 'xl'])
thumbnails.sort(key=lambda thumb: QUALITIES.index(thumb[1]) if thumb[1] in QUALITIES else -1) largest = max(thumbnails, key=lambda thumb: quality_key(thumb[1]))
thumbnail = 'http://www.ndr.de' + thumbnails[-1][0] thumbnail = 'http://www.ndr.de' + largest[0]
for format_id in ['lo', 'hi', 'hq']: for format_id in 'lo', 'hi', 'hq':
formats.append({ formats.append({
'url': '%s.%s.mp4' % (video_url.group('video'), format_id), 'url': '%s.%s.mp4' % (video_url.group('video'), format_id),
'format_id': format_id, 'format_id': format_id,