PublicMirrors/you-get
1
0
Fork 0
mirror of https://github.com/soimort/you-get.git synced 2025-01-23 21:45:02 +03:00
Code Issues Actions Packages Projects Releases 1 Wiki Activity

[tumblr] filter out some non-image urls

Browse Source
This commit is contained in:
Mort Yao 2015-10-29 20:47:18 +01:00
parent 42ed56d2fd
commit 214deb6c8b
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/you_get/extractors/tumblr.py
View File

@ -14,9 +14,9 @@ def tumblr_download(url, output_dir='.', merge=True, info_only=False, **kwargs):
page_title = r1(r'<meta name="description" content="([^"\n]+)', html) or \ page_title = r1(r'<meta name="description" content="([^"\n]+)', html) or \
r1(r'<meta property="og:description" content="([^"\n]+)', html) or \ r1(r'<meta property="og:description" content="([^"\n]+)', html) or \
r1(r'<title>([^<\n]*)', html) r1(r'<title>([^<\n]*)', html)
urls = re.findall(r'(https?://[^;"]+/tumblr_[^;"]+_\d+\.jpg)', html) +\ urls = re.findall(r'(https?://[^;"&]+/tumblr_[^;"]+_\d+\.jpg)', html) +\
re.findall(r'(https?://[^;"]+/tumblr_[^;"]+_\d+\.png)', html) +\ re.findall(r'(https?://[^;"&]+/tumblr_[^;"]+_\d+\.png)', html) +\
re.findall(r'(https?://[^;"]+/tumblr_[^";]+_\d+\.gif)', html) re.findall(r'(https?://[^;"&]+/tumblr_[^";]+_\d+\.gif)', html)
tuggles = {} tuggles = {}
for url in urls: for url in urls: