Skip to content
Silver edited this page Jan 11, 2023 · 30 revisions

Table of Contents

Introduction

WikiExtractor.py is a script that extracts and cleans text from a Wikipedia database dump.

The tool is written in Python and requires no additional library.

Wikipedia articles are written in the MediaWiki Markup Language which provides a simple notation for formatting text (bolds, italics, underlines, images, tables, etc.). It also allows inserting HTML markup in the documents. Wiki and HTML tags are sometimes misused (unclosed tags, wrong attributes, etc.), therefore the extractor deploys some heuristics in order to circumvent such problems.

WikiExtractor.py is capable of performing template expansion to some extent: it does not fully support Lua modules.

Usage

The script is invoked with a Wikipedia dump file as an argument. Use the article dumps which are available as http://dumps.wikimedia.org/XXwiki/latest/XXwiki-latest-pages-articles.xml.bz2, where XX is the language identifier (e.g. en, es, zh).

The output is stored in a number of files of similar size in a chosen directory. Each file will contains several documents in this file format.

Template expansion requires preprocessing first the whole dump and collecting template definitions.

Usage:

  python -m wikiextractor.WikiExtractor [options] xml-dump-file

optional arguments:

  -h, --help            show this help message and exit
  -o OUTPUT, --output OUTPUT
                        directory for extracted files (or '-' for dumping to
                        stdout)
  --processes PROCESSES
                        Number of processes to use (default 23)
  -b n[KMG], --bytes n[KMG]
                        put specified bytes per output file (default is 1M)
  -c, --compress        compress output files using bzip
  -l, --links           preserve links
  -ns ns1,ns2, --namespaces ns1,ns2
                        accepted namespaces
  -q, --quiet           suppress reporting progress info
  --debug               print debug info
  -s, --sections        preserve sections
  -a, --article         analyze a file containing a single article
  --templates TEMPLATES
                        use or create file containing templates
  --no_templates        do not expand templates
  -r, --revision        Include the document revision id (default=False)
  --min_text_length MIN_TEXT_LENGTH
                        Minimum expanded text length required to write
                        document (default=0)
  --filter_disambig_pages
                        Remove pages from output that contain disabmiguation
                        markup (default=False)
  --threads THREADS     Number of threads to use (default 8)
  -v, --version         print program version

Saving templates to a file will speed up performing extraction the next time, assuming template definitions have not changed.

Option --no-templates significantly speeds up the extractor, avoiding the cost of expanding MediaWiki templates.

Wikipedia dumps

Related Work

Clone this wiki locally