Importing Mediawiki Content into Dokuwiki

Beware, all shell code listings here are for zsh on Debian!

You need:

• A working Dokuwiki instance
• API access to your mediawiki
• For importing users: Access to the database of your Mediawiki installation
• yamdwe: Github
  • As of 2018-11-05, using a MariaDB 10.2 and yamdwe with a virtualenv, I had to install libmariadb-dev and libmariadb-dev-compat in version 2.* to compile MySQL-python

1. Import pages and media using yamdwe.py
2. Import users using yamdwe_users.py
3. cd into your Dokuwiki installation directory
4. Add page titles from the page file names:

   for f in $(find data/pages/ -maxdepth 1 -type f) ; do
       t=$(echo $f | sed -E -e 's#^.+/(.+)\.txt$#\1#' -e 's#_# #g' -e 's#\b.#\u&#g')
       cp "$f" "$f.backup"
       sed -i "1i====== $t ======\\n" $f
   done

   1. I only had pages in the root namespace, find data/pages/ -maxdepth 1 -type f finds all those
   2. The first sed builds the page title and uses extended regex (-E):
      1. s#^.+/(.+)\.txt$#\1# strips .txt from the file name
      2. s#_# #g replaces underscores with spaces
      3. s#\b.#\u&#g converts the string to Title Case
   3. The second sed inserts (i) ====== $t ======\\n as the first line (1)
5. Check your pages
6. Maybe fix the page file permissions and page titles (title case doesn't work well for most abbreviations…)
7. Delete backups: find data/pages -maxdepth 1 -type f -name \*.txt.backup -execdir rm \{\} \;
8. yamdwe inserted a comment at the bottom of the imported mediawiki main page, delete that maybe

# Install Dokuwiki
# Install libmariadb-dev and libmariadb-dev-compat in version 2.*
> git clone https://github.com/projectgus/yamdwe.git
> cd yamdwe
> virtualenv -p python2.7 env
> pip install -r requirements.txt
> python yamdwe.py https://mediawiki.example.org/api.php path/to/dokuwiki
> touch path/to/dokuwiki/conf/local.php
> python yamdwe_users.py -u mediawiki --db wiki path/to/dokuwiki
> for f in $(find data/pages/ -maxdepth 1 -type f) ; do t=$(echo $f | sed -E -e 's#^.+/(.+)\.txt$#\1#' -e 's#_# #g' -e 's#\b.#\u&#g'); cp "$f" "$f.backup" ; sed -i "1i======$t======\\n" $f ; done
# Check your pages!
> find data/pages -maxdepth 1 -type f -name \*.txt.backup -execdir rm \{\} \;