WikiFundi is a solution which provide a pre-configured Mediawiki in a similar way like Wikipedia. Probably the easiest way to train users if you do not have access to Internet!

Here steps to install it with Docker or on a RaspberryPi.

To build and run :

mkdir -p data
docker build -t wikifundi_en .
docker run -p 8080:80 -v ${PWD}/data:/var/www/data -it wikifundi_en

Go to http://localhost:8080/

Default admin logging :

• User : Admin
• Password : wikiadmin

The WikiFundi image extends the openzim/mediawiki Docker image to allow mirroring a existing wiki (by example Wikipedia) and use this wiki offline

The config files are located in config directory.

This directory contain the configuration of :

• mirroring/mirroring.json : Pages to copy from an other wiki and modifications after copy. To get file structure : export PYWIKIBOT2_DIR=config/pywikibot/` ./wikimedia_sync.py --help
• mediawiki/LocalSettings.custom.php : You can customise the Mediawiki by editing your this file. If you want to know more, have a look to documentation
• parsoid/config.yaml : Parsoid config file (allow to use VisualEdit)
• pywikibot/user-config.py : Configure pywikibot library to use MediaWiki API (needed for mirroring)

You can also customize your logo in assets/images as customize the assets/docs/README embeded with data

On start, if the database not exist, then it is initialized. You can lauch mirroring by changing environments variables MIRRORING :

docker run -p 8080:80 -e MIRRORING=1 -v ${PWD}/data:/var/www/data -it wikifundi_en

You can also change options script with MIRRORING_OPTIONS :

• -f, --force : always copy the content (even if page exist on site dest) (default : false)
• -t, --no-sync-templates : do not copy templates used by the pages to sync. (default : false)
• -u, --no-upload-files : do not copy files (images, css, js, sounds, ...) used by the pages to sync (default : false)
• -p, --no-sync : do not copy anything (default : false)
• -m, --no-modify : do not modify pages (default : false)
• -r, --resume : try to resume previous sync (default : false)
• -e, --export-dir : write resume files in this directory (default : current directory)
• -w, --thumbwidth :try to download thumbnail image with this width instead original image (default : 1024)
• -s, --maxsize : do not files download greater to this limit (default : 100MB)
• -a, --async : execute mirroring in async mode (5 threads / cpu). No works with SQLITE database. (default : false)

Case usaging : ./wikimedia_sync.py -m 5MB -w 2000 config.json : sync page, templates, files and modify pages. Do not copy file > 5MB and Copy images (jpeg and png) in 2000px (if available). ./wikimedia_sync.py -tu config.json : sync and modify pages. Do not copy dependances (templates and files). ./wikimedia_sync.py -p config.json : just modify pages. ./wikimedia_sync.py -pm config.json : do anything. ./wikimedia_sync.py -af config.json : copy all pages and their dependencies in async mode.

After mirroring, you can generate tarball by going http://localhost:8080/export_data.php