Bugs item #3536400, was opened at 2012-06-19 12:52
Message generated for change (Comment added) made by xqt
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=603138&aid=353640…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: betacommand (betacommand)
Assigned to: xqt (xqt)
Summary: Invalid Title in flickrripper
Initial Comment:
Betacommand multichill: I know you wrote flickrripper.py and Im trying to fix an issue
with it, and thought it might be easier for you to fix
Betacommand lines 157-161 where it grabs the description and uses it for the file name
Betacommand when you start working with non-latin descriptions it doesnt handle multi-byte
characters well, it ended up with a title over 320 bytes
Betacommand the max mediawiki lets you have is 255
multichill Lol
Betacommand multichill: really rather a pain
multichill So the check shoul probably encode it and than see how long it is?
Betacommand correct
multichill Or just lower the limit a bit?
Betacommand thai letters for example are 3 bytes
Betacommand notes it was discovered with flickrripper.py -autonomous -user_id:40561337@N07
-addcategory:"Files from Abhisit Vejjajiva Flickr stream"
multichill Betacommand: Could you file a bug for this?
Betacommand multichill: you would need to cut it down to 85 to be safe
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-20 23:20
Message:
Thanks for testing. The lenght calculation was wrong. I've corrected it
----------------------------------------------------------------------
Comment By: betacommand (betacommand)
Date: 2012-06-20 22:04
Message:
Line 787 doesnt return the title, it returns the whole page (self) when you
print the object and not the title it gets converted there. I used a log to
confirm that the title was UTF-8 before filling this bug,
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-20 21:48
Message:
I do not see a conversion by the exception. I converted the title from html
entities to unicode in my last commit
----------------------------------------------------------------------
Comment By: betacommand (betacommand)
Date: 2012-06-20 20:55
Message:
those are the thai parts of the page title that are being converted when
the exception is being thrown
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-20 20:53
Message:
Where are the html entities from? Are they part of the flickr page?
----------------------------------------------------------------------
Comment By: betacommand (betacommand)
Date: 2012-06-20 16:45
Message:
Issue still not fixed, actually its worse
C:\Dev\SVN\pywikipedia>flickrripper.py -autonomous -user_id:40561337@N07
-addcat
egory:"Files from Abhisit Vejjajiva Flickr stream"
5703017392
Traceback (most recent call last):
File "C:\Dev\SVN\pywikipedia\flickrripper.py", line 609, in
<module>
main()
File "C:\Dev\SVN\pywikipedia\flickrripper.py", line 599, in main
removeCategories, autonomous)
File "C:\Dev\SVN\pywikipedia\flickrripper.py", line 257, in
processPhoto
filename = getFilename(photoInfo)
File "C:\Dev\SVN\pywikipedia\flickrripper.py", line 172, in
getFilename
% (title, project, username)).exists():
File "C:\Dev\SVN\pywikipedia\wikipedia.py", line 1284, in exists
self.get()
File "C:\Dev\SVN\pywikipedia\wikipedia.py", line 705, in get
expandtemplates = expandtemplates)
File "C:\Dev\SVN\pywikipedia\wikipedia.py", line 787, in
_getEditPage
raise BadTitle('BadTitle: %s' % self)
pywikibot.exceptions.BadTitle: BadTitle:
[[commons:File:นาย
585;รัฐมนตรี
แลū
2;คณะเดินทางอ&
#3629;กจากกรุงจา#
85;าร์ตา
สาธารณ
;รัฐอินโดนีเ&#
3595;ียกลับยังปũ
9;ะเทศไทย
วันอ
าทิตย์ที่
8
พฤ
ษภาคม พ.ศ.2554
(Photographer
attached
to the Prime Minister of the Kingdom of Thailand (H.E.Mr.Abhisit
Vejjajiva) , Pe
erapat Wimolrungkarat -
พีรพัฒน์
วิมลรังครัต&#
3609;์) @is50mm - Flickr - Abhisit Vejjajiva.jpg]]
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-20 07:17
Message:
fix committed in r10387, please check
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-20 00:08
Message:
an idea for getFilename (could anybody test it whether it works)
if not title:
#find the max length for a mw title
maxBytes = 240 - len(project.encode('utf-8')) \
- len(username.encode('utf-8'))
description = photoInfo.find('photo').find('description').text
if description:
descBytes = len(description.encode('utf-8'))
if descBytes > maxBytes:
# maybe we cut more than needed, anyway we do it
items = max(0, len(description) - maxBytes + descBytes)
description = description[:items]
title = cleanUpTitle(description)
else:
title = u''
# Should probably have the id of the photo as last resort.
----------------------------------------------------------------------
Comment By: xqt (xqt)
Date: 2012-06-19 23:03
Message:
I guess the title is cutted by mw and not the slice operator since it works
correct for unicode strings. len() also gives the number of characters not
the number bytes. Do we have any size(object) method?
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=603138&aid=353640…