Benutzer:Dirk Hünniger/wb2pdf
Summary
[Bearbeiten]mediawiki2latex converts MediaWiki markup to LaTeX and, via LaTeX, to PDF. It can be used to export pages from any project running MediaWiki, such as Wikipedia. It is also possible to generate EPUB and ODT output files.
Web Version
[Bearbeiten]You may test mediawiki2latex here:
https://mediawiki2latex.wmcloud.org/
Remember:
- There is a time limit of four hours ( 2000 pages) per request on the server.
- There is no limit on the locally installed versions described below.
Installation Instructions
[Bearbeiten]User Manual
[Bearbeiten]see the User Manual
Command Line Version
[Bearbeiten]A command line version is currently available as part of the Stretch Debian distribution, as well as the current Ubuntu distribution.
LaTeX intermediate Code
[Bearbeiten]On Linux you can use the -c command line option with an absolute path name.
Media
[Bearbeiten]Talk
[Bearbeiten]English: File:Wb2pdfTalk.ogv
German (Deutsch): File:MediaWiki2LaTeXTalkRevision4.webm
Slides
[Bearbeiten]English: File:Wb2pdfTalk.pdf
German (Deutsch): File:MediaWiki2LaTeXTalk.pdf
Poster
[Bearbeiten]
In Action
[Bearbeiten]To see it in action look here: Datei:Wb2latexCompilingWikibook2PDF.ogg
Developers
[Bearbeiten]The follwing Link Benutzer:Dirk Huenniger/wb2pdf/details explains some of the inner workings of the software.
Quality and Statistics
[Bearbeiten]A test run in October 2014 processing 4369 featured articles of the English Wikipedia did produce a PDF file in each case. In particular these were all featured articles we were able to find at the beginning of the test.
In May 2020 we looked at the usage of the web server an saw that the 50 requests looked at resulted in the following output:
| 37 | |
| FAIL | 2 |
| EPUB | 3 |
| ZIP | 2 |
| ODT | 6 |
The failures are believed to be caused by trying to process large books from the wikipedia book namespace exceeding the time limit of four hours.
In December 2018 we also did a test run on 100 featured articles on the English wikipedia. In two cases a PDF was not created. We just ran these two cases once again and got a PDF in each case. The total size of all PDFs was 2.2 GB on disk. 5 GB were downloaded in order to make them. The process took 6 hours and 15 minutes. The computer used was a i5-8250U notebook with 8 GByte of memory running only one instance of mediawiki2latex at a time. The internet downstream speed was 11.6 MBit/s.
Currently, the largest book we've created with mediawiki2latex is 8991 pages long.