Jump to navigation Jump to search
Git logo
Wikimedia Git repository has this file: scripts/ is a script which lets you easily add pages to a category of your choosing, based upon an investigation of links in existing articles. Its potential applications are diverse, but it would be especially helpful on wikis that do not yet have well-established category trees. Used properly, it could create a number of fully populated categories much more quickly than by manual editing.

This script is a standard part of the pywikipediabot installation, and should be available to all users who have installed pywikipediabot properly. However, in most installations of the pywikipedia script library, it produces syntax errors and immediately aborts. These errors have been present in the script's released version since at least 4Q 2009. They are, however, easily corrected. Although it does not work "out of the box" as of January 2012, you can get it up and running in less than one minute.


The basic idea of the script is straightforward, though it relies on an assumption that may not be appropriate for all wikis. It allows you to use the links on a page to quickly populate a category of your choosing. It presumes, therefore, that pages about a certain topic will have links that you want to put into a category about the same topic.

For instance, imagine you have a page called American football and you want to quickly populate category:American football teams. The script will go through the text of the page, stopping on everything included in link brackets. It will then ask you whether you want to put that linked page into the category:American football teams. Assuming that American football will contain links to Wahington Redskins, Pittsburgh Steelers and other teams would be linked on a page about American football, would thus save you a lot of time over going to each one of those team pages and adding a category manually..


The script is invoked by python Without any parameters, this will prompt you for a page to start with, and you'll then go from there, depending on which pages you admit to your category. Despite the ease of this arrangement, you'll probably want to add some parameters. If you give no parameters, it assumes you want to add the discovered pages in whatever category you were working on the last time you used the bot.

A more specific syntax is: python "category name" -argument

Note that category name should just be the category name, with namespace and brackets excluded. Technically, the enclosing quotation marks aren't necessary, but they are allowed. The argument requires a prepending hyphen and can be one of a number of things, as seen in the script's standard help file:

This bot takes as its argument (or, if no argument is given, asks for it), the
name of a new or existing category. It will then try to find new articles for
this category (pages linked to and from pages already in the category), asking
the user which pages to include and which not.

   -nodates  automatically skip all pages that are years or dates (years
             only work AD, dates only for certain languages)
   -forward  only check pages linked from pages already in the category,
             not pages linking to them. Is less precise but quite a bit
   -exist    only ask about pages that do actually exist; drop any
             titles of non-existing pages silently. If -forward is chosen,
             -exist is automatically implied.
   -keepparent  do not remove parent categories of the category to be
             worked on.
   -all      work on all pages (default: only main namespace)

When running the bot, you will get one by one a number by pages. You can
Y(es) - include the page
N(o) - do not include the page or
I(gnore) - do not include the page, but if you meet it again, ask again.
X - add the page, but do not check links to and from it
Other possiblities:
A(dd) - add another page, which may have been one that was included before
C(heck) - check links to and from the page, but do not add the page itself
R(emove) - remove a page that is already in the list
L(ist) - show current list of pages to include or to check

So, to follow with our American football example, python "American football teams" -keepparent -forward will do the following:

  1. Assume that the page to start on is itself titled "American football teams", but give you the opportunity to choose a different starting page
  2. Make sure that the category tree is undisturbed by keeping the parent of category:American football teams, perhaps category:Sporting teams, in place. (You should give strong consideration to almost always using the -keepparent argument.)
  3. It will only search for pages which exist and are linked to on your page — in other words, pages that are actually forward of the one on which you're working

As the bot works and you accept certain pages into your new category, it will then branch off and look from the pages that you add to the category. So, imagine you find a link to Dallas Cowboys on American football. You accept that as a page that needs to be in category:American football teams. The script will now also scan Dallas Cowboys for links. It's this additive process that means you'll more quickly amass pages in your desired category.

Not automated[edit]

There are no circumstances in which a reasonable wiki administrator would ever want to automate this script, so no option for doing so is included. Think of this as a "semi-automatic" script. It's still much faster than just adding individual pages to a category, but it doesn't work without human control of the bot.

Error correction[edit]

The version which is part of the standard pywikipedia script library as of January 2012 is 7336, dated 2009-09-29. This code has a small syntax error that needs to be corrected before the script can be used.

  1. Navigate to within your library. (On Macs, this is typically at whatever_your_user_name_is/pywikipedia/
  2. Open by double-clicking on it. Unless you have a very barebones setup, your computer should automatically find a program capable of editing it. (On Macs, IDLE should instantly start on double click.)
  3. Navigate down to the area just below the explanatory text at the top of the file. You'll see a raft of different translations of a one-line message. In English, this message says "creation or update of category". (If you're lost, just use your editor's search function to look for that phrase.)
  4. If you know a bit about javascript coding, you'll be able to instantly see some glaring syntax errors at the end of the 'nn' and 'no' lines. 'nn':u'oppretting eller oppdatering av kategori:': and 'no':u'opprettelse eller oppdatering av kategori:': are the source of the problem.
  5. Change the final colons to commas. The lines should read: 'nn':u'oppretting eller oppdatering av kategori:', and 'no':u'opprettelse eller oppdatering av kategori:',
  6. Save your file (Ctrl-S; or, if Mac, Command-S) will now work properly..