Skip to main content

A simple parsing tool for MediaWiki's wikitext markup.

Project description


A simple to use WikiText parsing library for MediaWiki.

The purpose is to allow users easily extract and/or manipulate templates, template parameters, parser functions, tables, external links, wikilinks, lists, etc. found in wikitexts.


  • Python 3.4+ is required
  • pip install 'setuptools>=36.2.1'
  • pip install wikitextparser


>>> import wikitextparser as wtp

WikiTextParser can detect sections, parser functions, templates, wiki links, external links, arguments, tables, wiki lists, and comments in your wikitext. The following sections are quick of some of these functionalities.

You may also want to have a look at the test modules for more examples and probable pitfalls (expected failures).


>>> parsed = wtp.parse("{{text|value1{{text|value2}}}}")
>>> parsed.templates
[Template('{{text|value1{{text|value2}}}}'), Template('{{text|value2}}')]
>>> parsed.templates[0].arguments
>>> parsed.templates[0].arguments[0].value = 'value3'
>>> print(parsed)

The pformat method returns a pretty-print formatted string for templates:

>>> parsed = wtp.parse('{{t1 |b=b|c=c| d={{t2|e=e|f=f}} }}')
>>> t1, t2 = parsed.templates
>>> print(t2.pformat())
    | e = e
    | f = f
>>> print(t1.pformat())
    | b = b
    | c = c
    | d = {{t2
        | e = e
        | f = f

Template.rm_dup_args_safe and Template.rm_first_of_dup_args methods can be used to clean-up pages using duplicate arguments in template calls:

>>> t = wtp.Template('{{t|a=a|a=b|a=a}}')
>>> t.rm_dup_args_safe()
>>> t
>>> t = wtp.Template('{{t|a=a|a=b|a=a}}')
>>> t.rm_first_of_dup_args()
>>> t

Template parameters:

>>> param = wtp.parse('{{{a|b}}}').parameters[0]
>>> param.default
>>> param.default = 'c'
>>> param
>>> param.append_default('d')
>>> param


>>> parsed = wtp.parse("""
... == h2 ==
... t2
... === h3 ===
... t3
... === h3 ===
... t3
... == h22 ==
... t22
... {{text|value3}}
... [[Z|X]]
... """)
>>> parsed.sections
 Section('== h2 ==\nt2\n=== h3 ===\nt3\n=== h3 ===\nt3\n'),
 Section('=== h3 ===\nt3\n'),
 Section('=== h3 ===\nt3\n'),
 Section('== h22 ==\nt22\n{{text|value3}}\n[[Z|X]]\n')]
>>> parsed.sections[1].title = 'newtitle'
>>> print(parsed)

=== h3 ===
=== h3 ===
== h22 ==
>>> del parsed.sections[1].title
>>>> print(parsed)

=== h3 ===
=== h3 ===
== h22 ==


Extracting cell values of a table:

>>> p = wtp.parse("""{|
... |  Orange    ||   Apple   ||   more
... |-
... |   Bread    ||   Pie     ||   more
... |-
... |   Butter   || Ice cream ||  and more
... |}""")
>>> p.tables[0].data()
[['Orange', 'Apple', 'more'],
 ['Bread', 'Pie', 'more'],
 ['Butter', 'Ice cream', 'and more']]

By default, values are arranged according to colspan and rowspan attributes:

>>> t = wtp.Table("""{| class="wikitable sortable"
... |-
... ! a !! b !! c
... |-
... !colspan = "2" | d || e
... |-
... |}""")
[['a', 'b', 'c'], ['d', 'd', 'e']]
[['a', 'b', 'c'], ['d', 'e']]

Calling the cells method of a Table returns table cells as Cell objects. Cell objects provide methods for getting or setting each cell’s attributes or values individually:

>>> cell = t.cells(row=1, column=1)
>>> cell.attrs
{'colspan': '2'}
>>> cell.set('colspan', '3')
>>> print(t)
{| class="wikitable sortable"
! a !! b !! c
!colspan = "3" | d || e

HTML attributes of Table, Cell, and Tag objects are accessible via get_attr, set_attr, has_attr, and del_attr methods.


The get_lists method provides access to lists within the wikitext.

>>> parsed = wtp.parse(
...     'text\n'
...     '* list item a\n'
...     '* list item b\n'
...     '** sub-list of b\n'
...     '* list item c\n'
...     '** sub-list of b\n'
...     'text'
... )
>>> wikilist = parsed.get_lists()[0]
>>> wikilist.items
[' list item a', ' list item b', ' list item c']

The sublists method can be used to get all sub-lists of the current list or just sub-lists of specific items:

>>> wikilist.sublists()
[WikiList('** sub-list of b\n'), WikiList('** sub-list of b\n')]
>>> wikilist.sublists(1)[0].items
[' sub-list of b']

It also has an optional pattern argument that works similar to lists, except that the current list pattern will be automatically added to it as a prefix:

>>> wikilist = wtp.WikiList('#a\n#b\n##ba\n#*bb\n#:bc\n#c', '\#')
>>> wikilist.sublists()
[WikiList('##ba\n'), WikiList('#*bb\n'), WikiList('#:bc\n')]
>>> wikilist.sublists(pattern='\*')

Convert one type of list to another using the convert method. Specifying the starting pattern of the desired lists can facilitate finding them and improves the performance:

>>> wl = wtp.WikiList(
...     ':*A1\n:*#B1\n:*#B2\n:*:continuing A1\n:*A2',
...     pattern=':\*'
... )
>>> print(wl)
:*:continuing A1
>>> wl.convert('#')
>>> print(wl)
#:continuing A1


Accessing HTML tags:

>>> p = wtp.parse('text<ref name="c">citation</ref>\n<references/>')
>>> ref, references = p.get_tags()
>>> = 'X'
>>> ref
Tag('<X name="c">citation</X>')
>>> references

WikiTextParser is able to handle common usages of HTML and extension tags. However it is not a fully-fledged HTML parser and may fail on edge cases or malformed HTML input. Please open an issue on github if you encounter bugs.


parent and ancestors methods can be used to access a node’s parent or ancestors respectively:

>>> template_d = parse("{{a|{{b|{{c|{{d}}}}}}}}").templates[3]
>>> template_d.ancestors()
>>> template_d.parent()
>>> _.parent()
>>> _.parent()
>>> _.parent()  # Returns None

Use the optional type_ argument if looking for ancestors of a specific type:

>>> parsed = parse('{{a|{{#if:{{b{{c<!---->}}}}}}}}')
>>> comment = parsed.comments[0]
>>> comment.ancestors(type_='ParserFunction')

To delete/remove any object from its parents use del object[:] or del object.string.

Compared with mwparserfromhell

mwparserfromhell is a mature and widely used library with nearly the same purposes as wikitextparser. The main reason leading me to create wikitextparser was that mwparserfromhell could not parse wikitext in certain situations that I needed it for. See mwparserfromhell’s issues 40, 42, 88, and other related issues. In many of those situation wikitextparser may be able to give you more acceptable results.

But if you need to

then mwparserfromhell or maybe other libraries will be the way to go. Also note that wikitextparser is still under heavy development and the API may change drastically in the future versions.

Of-course wikitextparser has its own unique features, too: Providing access to individual cells of each table, pretty-printing templates, and a few other advanced functions.

The tokenizer in mwparserfromhell is written in C. Tokenization in wikitextparser is mostly done using the regex library which is also in C. I have not rigorously compared the two libraries in terms of performance, i.e. execution time and memory usage. In my limited experience, wikitextparser has a decent performance and should able to compete and may even have little performance benefits in many situations. However if you are working with on-line data, any difference is usually negligible as the main bottleneck will be the network latency.

If you have had a chance to compare these libraries in terms of performance please share your experience by opening an issue on github.

Known issues and limitations

  • Syntax elements produced by a template transclusion cannot be detected by offline parsers.
  • Localized namespace names are unknown, so for example [[File:...]] links are treated as normal wikilinks. mwparserfromhell has similar issue, see #87 and #136. As a workaround, Pywikibot can be used for determining the namespace.
  • Linktrails are language dependant and are not supported. Also not supported by mwparserfromhell. However given the trail pattern and knowing that wikilink.span[1] is the ending position of a wikilink, it should be trivial to compute a WikiLink’s linktrail.
  • Templates adjacent to external links, are never considered part of the link. In reality, this depends on the contents of the template. Example: parse('{{dead link}}').external_links[0].url == ''
  • While MediaWiki recognizes only a finite number of tags and they are extension-dependent, the tags method returns anything that looks like an HTML tag. A configuration option might be added in the future to address this issue.

Project details

Release history Release notifications

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for wikitextparser, version 0.34.0
Filename, size File type Python version Upload date Hashes
Filename, size wikitextparser-0.34.0-py3-none-any.whl (60.2 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size wikitextparser-0.34.0.tar.gz (58.8 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page