mwlib

The library is intended for programmers or developers working on wiki systems supporting the MediaWiki synthax or web-designers parsing and transforming content from Wikipedia.

The library is only for parsing MediaWiki, outputting in various formats is done via extensions.

mwlib can render articles to:
mwxml
XHTML 1.0 Transitional
DocBook XML
OpenDocument (Text)
PDF
epub
ZIM
CDB

What is new in this release:

• Workaround ‘first run' tox issue.
• Fix tests with new wsgi_intercept 0.6 and require that version.
• Match IPv6 addresses as anonymous users.

What is new in version 0.15.11:

• Fixed xnet tests
• Uses os.urandom in utils.uid
• Generated junitxml files
• Removed empty reference nodes

What is new in version 0.15.8:

• Also fetch used images when fetching ‘redirected revisions'.

What is new in version 0.15.7:

• Removed explicitly positioned nodes regardless of nesting level. fix bug where children were skipped and not removed.

What is new in version 0.15.1:

• Fixed mw-serve-ctl.

What is new in version 0.14.0:

• Install scripts via plain old distutils instead of "console_scripts" entry point
• Remove cdbwiki
• Remove mwlib.xfail, use pytest.mark.xfail instead
• Expect setuptools or distribute to be installed

What is new in version 0.12.16:

• Removed xhtmlwriter.
• Removed docbookwriter.
• fix_wikipedia_siteinfo for kdb, ltg and xmf.
• Removed zipwiki.
• Implement safesubst.
• Match noinclude and onlyinclude tags with whitespace.
• Bail out when running setup.py with an unsupported python version.

What is new in version 0.12.15:

• Require lxml.
• Don't switch fonts for direction switch chars lrm/rlm.
• Set teletype style by CSS.
• Fix rtl direction check bug.
• Quick fix in order to support the kbd tag.
• Fixed switch statements with localized #default case.
• Don't remove direction switching nodes.
• Resolve aliases when expanding templates.
• Support localized parser functions.
• Make tests work with latest py.test 2.1.
• Added support for css direction switching.
• Code and Var nodes now use teletype style.
• Be more verbose when collection params can not be retrieved.
• get rid of some more parsing calls.
• Cache img display info in licensehandler.
• Always show full text of contributors of images.
• Fix for getAllDisplayText.
• Added nofilter to licensehandling.
• Make licensechecker less fragile to bad config format.
• Improve image license handling.
• Decrease api_request_limit.
• Limit max. simultaneous img downloads to 15.
• Increase standard resolution of images.
• Fix getting html with revisions.
• Clean up after fixNesting.
• Fetch extension images.
• Prevent adding same API URL twice.
• Retry failed img downloads.
• Workaround for missing descriptionurl.
• Support older mediawikis.

What is new in version 0.12.14:

• Fix NS magic function.
• Do not parse links if link target would contain newlines.
• Removed collapsible elements
• Move gallery nodes out of tables.