Administrative scripts#
blockpageschecker script#
A bot to remove stale protection templates from pages that are not protected
Very often sysops block the pages for a set time but then they forget to remove the warning! This script is useful if you want to remove those useless warning left in these pages.
These command line parameters can be used to specify which pages to work on:
This script supports use of pagegenerators arguments.
Furthermore, the following command line parameters are supported:
-protectedpages Check all the blocked pages; useful when you have not
categories or when you have problems with them. (add the
namespace after ":" where you want to check - default checks
all protected pages.)
-moveprotected Same as -protectedpages, for moveprotected pages
This script is a ConfigParserBot.
The following options can be set within a settings file which is scripts.ini
by default::
-always Doesn't ask every time whether the bot should make the change.
Do it always.
-show When the bot can't delete the template from the page (wrong
regex or something like that) it will ask you if it should
show the page on your browser.
(attention: pages included may give false positives!)
-move The bot will check if the page is blocked also for the move
option, not only for edit
Examples::
python pwb.py blockpageschecker -always
python pwb.py blockpageschecker -cat:Geography -always
python pwb.py blockpageschecker -show -protectedpages:4
delete script#
This script can be used to delete and undelete pages en masse
Of course, you will need an admin account on the relevant wiki.
These command line parameters can be used to specify which pages to work on:
This script supports use of pagegenerators arguments.
Furthermore, the following command line parameters are supported:
-always Don't prompt to delete pages, just do it.
-summary:XYZ Set the summary message text for the edit to XYZ.
-undelete Actually undelete pages instead of deleting.
Obviously makes sense only with -page and -file.
-isorphan Alert if there are pages that link to page to be
deleted (check 'What links here').
By default it is active and only the summary per namespace
is be given.
If given as -isorphan:n, n pages per namespace will be shown,
If given as -isorphan:0, only the summary per namespace will
be shown,
If given as -isorphan:n, with n < 0, the option is disabled.
This option is disregarded if -always is set.
-orphansonly: Specified namespaces. Separate multiple namespace
numbers or names with commas.
Examples:
-orphansonly:0,2,4
-orphansonly:Help,MediaWiki
Note that Main ns can be indicated either with a 0 or a ',':
-orphansonly:0,1
-orphansonly:,Talk
Usage:
python pwb.py delete [-category categoryName]
Examples
Delete everything in the category “To delete” without prompting:
python pwb.py delete -cat:"To delete" -always
patrol script#
The bot is meant to mark the edits based on info obtained by whitelist
This bot obtains a list of recent changes and newpages and marks the edits as patrolled based on a whitelist.
Whitelist Format#
The whitelist is formatted as a number of list entries. Any links outside of lists are ignored and can be used for documentation. In a list the first link must be to the username which should be white listed and any other link following is adding that page to the white list of that username. If the user edited a page on their white list it gets patrolled. It will also patrol pages which start with the mentioned link (e.g. [[foo]] will also patrol [[foobar]]).
To avoid redlinks it’s possible to use Special:PrefixIndex as a prefix so that it will list all pages which will be patrolled. The page after the slash will be used then.
On Wikisource, it’ll also check if the page is on the author namespace in which case it’ll also patrol pages which are linked from that page.
An example can be found at https://en.wikisource.org/wiki/User:Wikisource-bot/patrol_whitelist
Commandline parameters:
-namespace Filter the page generator to only yield pages in
specified namespaces
-ask If True, confirm each patrol action
-whitelist page title for whitelist (optional)
-autopatroluserns Takes user consent to automatically patrol
-versionchecktime Check versionchecktime lapse in sec
-repeat Repeat run after 60 seconds
-newpages Run on unpatrolled new pages
(default for Wikipedia Projects)
-recentchanges Run on complete unpatrolled recentchanges
(default for any project except Wikipedia Projects)
-usercontribs Filter generators above to the given user
protect script#
This script can be used to protect and unprotect pages en masse
Of course, you will need an admin account on the relevant wiki. These command line parameters can be used to specify which pages to work on:
This script supports use of pagegenerators arguments.
Furthermore, the following command line parameters are supported:
-unprotect Acts like "default:all"
-default: Sets the default protection level (default 'sysop'). If no
level is defined it doesn't change unspecified levels.
-[type]:[level] Set [type] protection level to [level]
Usual values for [level] are: sysop, autoconfirmed, all; further levels may be provided by some wikis.
For all protection types (edit, move, etc.) it chooses the default protection level. This is “sysop” or “all” if -unprotect was selected. If multiple parameters -unprotect or -default are used, only the last occurrence is applied.
This script is a ConfigParserBot.
The following options can be set within a settings file which is scripts.ini
by default::
-always Don't prompt to protect pages, just do it.
-summary: Supply a custom edit summary. Tries to generate summary from
the page selector. If no summary is supplied or couldn't
determine one from the selector it'll ask for one.
-expiry: Supply a custom protection expiry, which defaults to
indefinite. Any string understandable by MediaWiki, including
relative and absolute, is acceptable. See:
:api:`Protect#Parameters`
Usage:
python pwb.py protect <OPTIONS>
Examples
Protect everything in the category ‘To protect’ prompting:
python pwb.py protect -cat:"To protect"
Unprotect all pages listed in text file ‘unprotect.txt’ without prompting:
python pwb.py protect -file:unprotect.txt -unprotect -always
revertbot script#
This script can be used for reverting certain edits
The following command line parameters are supported:
-username Edits of which user need to be reverted.
Default is bot's username (site.username())
-rollback Rollback edits instead of reverting them.
Note that in rollback, no diff would be shown.
-limit:num Use the last num contributions to be checked for revert.
Default is 500.
Users who want to customize the behaviour should subclass the BaseRevertBot
and override its callback method. Here is a sample:
class myRevertBot(BaseRevertBot)::
'''Example revert bot.'''
def callback(self, item) -> bool::
'''Sample callback function for 'private' revert bot.
:param item: an item from user contributions
:type item: dict
'''
if 'top' in item::
page = pywikibot.Page(self.site, item['title'])
text = page.get(get_redirect=True)
pattern = re.compile(r'\[\[.+?:.+?\..+?\]\]')
return bool(pattern.search(text))
return False
speedy_delete script#
Help sysops to quickly check and/or delete pages listed for speedy deletion
This bot trawls through candidates for speedy deletion in a fast and semi-automated fashion. It displays the contents of each page one at a time and provides a prompt for the user to skip or delete the page. Of course, this will require a sysop account.
Future upcoming options include the ability to untag a page as not being eligible for speedy deletion, as well as the option to commute its sentence to Proposed Deletion (see [[en:WP:PROD]] for more details). Also, if the article text is long, to prevent terminal spamming, it might be a good idea to truncate it just to the first so many bytes.
Warning
This tool shows the contents of the top revision only. It is possible that a vandal has replaced a perfectly good article with nonsense, which has subsequently been tagged by someone who didn’t realize it was previously a good article. The onus is on you to avoid making these mistakes.
Note
This script currently only works for the Wikipedia project.