This is my first post in English, so I apologize in advance for my poor grammar.
I’ll try to explain how to export all the pages of a Trac wiki into a Dokuwiki . All the process is based in my own shell script, which uses trac-admin, wget and html2wiki to export all the content of your Trac wiki into a dokuwiki folders.
At a glance, the main steps performed by the script are the following:
- Get the Trac wiki pages as a HTML files using
wget. - Remove all the HTML code belonging to the Trac page itself.
- Convert all the pages from HTML to Dokuwiki format (using
html2wiki). - Fix some internal/media links and transform footer attachments into dokuwiki format.
- Rename all the pages and media from upper to lower case.
The output of the script are two tarball files. One with the Dokuwiki pages, and the oter one with the media files.
As you probably know, the pages in a Dokuwiki are stored in simple text files,using the Dokuwiki format. All these pages are stored inside a folder called pages, using different subfolders for every namespace on the wiki. The same occour with the media files. All of them are stored inside folders named as the namespace they belong to, beneath the media folder. So, you just need to extract the content of the media tarball into the media folder and the content of the pages tarball inside the pages folder. And that’s all, your pages are now stored inside a Dokuwiki.
If you want to see the script, click on the Leer ms link below.
# Marcos Martinez
#
# Dependencies: html2wiki
# trac-admin
# wget
#
# Installation of html2wiki: perl -MCPAN -e 'install HTML::WikiConverter::DokuWiki'
#
TRAC_PROJECT_ENV='/path/to/trac/project'
TRAC_HOSTNAME='x.x.x.x or URL'
TRAC_PORT='80'
# Fill up the following two lines if your trac is using basic athentication method.
USER=your_trac_user
PASS=your_trac_password
# By default, all the pages will be exported inside a dokuwiki namespace named as the Trac Project
DOKUWIKI_NAMESPACE=$(grep '^name' $TRAC_PROJECT_ENV/conf/trac.ini | gawk '{print $3}')
#DOKUWIKI_NAMESPACE='your_desired_dokuwiki_namespace'
TRAC_BASE_NAME=$(grep 'base_url =' $TRAC_PROJECT_ENV/conf/trac.ini | gawk '{print $3}')
TRAC_URL_LOGIN=http://$TRAC_HOSTNAME:$TRAC_PORT$TRAC_BASE_NAME/login
TRAC_URL_WIKI=http://$TRAC_HOSTNAME:$TRAC_PORT$TRAC_BASE_NAME/wiki
TRAC_PROJECT_NAME=$(grep '^name' $TRAC_PROJECT_ENV/conf/trac.ini | gawk '{print $3}')
TRAC_INSTALL_DATE=$(ls -l $TRAC_PROJECT_ENV/README |gawk '{print $6 $7}')
TRAC_WIKI_PAGES=$(echo $TRAC_BASE_NAME/wiki | sed 's/\//\\\//g')
TRAC_WIKI_ATTACHMENTS=$(echo $TRAC_BASE_NAME/attachment/wiki| sed 's/\//\\\//g')
TRAC_WIKI_RAW_ATTACHMENTS=$(echo $TRAC_BASE_NAME/raw-attachment/wiki | sed 's/\//\\\//g')
TRAC_BS=$(echo $TRAC_BASE_NAME | sed 's/\//\\\//g')
# The PLIST var has the list of pages we want to export.
# By default, all the wiki pages modified after the installation date
# will be exported.
PLIST=$(trac-admin $TRAC_PROJECT_ENV wiki list|grep -v $TRAC_INSTALL_DATE | gawk '{print $1}'|sed '1,3d')
# Starting authenticated session
wget --quiet --spider \
--user=$USER --password=$PASS \
--save-cookies galletitas \
--keep-session-cookies $TRAC_URL_LOGIN
mkdir $DOKUWIKI_NAMESPACE
for PAGE in $PLIST
do
echo Downloading: $PAGE
wget --load-cookies galletitas --quiet -P $DOKUWIKI_NAMESPACE $TRAC_URL_WIKI/$PAGE
done
for FILE in `ls $DOKUWIKI_NAMESPACE`
do
echo Parsing $FILE
# Removing standard HTML trac code (header and footer)
sed -i '1,66d' $DOKUWIKI_NAMESPACE/$FILE
sed -i -n -e :a -e '1,45!{P;N;D;};N;ba' $DOKUWIKI_NAMESPACE/$FILE
# Removing table of contents (TOC) because in dokuwiki, the TOC is dynamically created at runtime.
sed -i '/class=\wiki-toc\/d' $DOKUWIKI_NAMESPACE/$FILE
# Adding HTML headers
sed -i '1i\html\nhead\n/head\nbody' $DOKUWIKI_NAMESPACE/$FILE
sed -i '$a\/body\n/html' $DOKUWIKI_NAMESPACE/$FILE
html2wiki --dialect DokuWiki $DOKUWIKI_NAMESPACE/$FILE $DOKUWIKI_NAMESPACE/$FILE.txt
rm $DOKUWIKI_NAMESPACE/$FILE
# Fixing internal links
sed -i s/\[\[$TRAC_WIKI_PAGES\//\[\[$DOKUWIKI_NAMESPACE:/g $DOKUWIKI_NAMESPACE/$FILE.txt
# Fixing media links
sed -i s/\[\[$TRAC_WIKI_ATTACHMENTS\/$FILE\//{{:$DOKUWIKI_NAMESPACE:/g $DOKUWIKI_NAMESPACE/$FILE.txt
sed -i s/{{$TRAC_WIKI_RAW_ATTACHMENTS\/$FILE\///g $DOKUWIKI_NAMESPACE/$FILE.txt
sed -i s/}}]]/}}/gDOKUWIKI_NAMESPACE/$FILE.txt
# Removing timeline data from 'Attachments' at the EOF if needed
NUM=$(grep -n '==== Attachments ====' $DOKUWIKI_NAMESPACE/$FILE.txt | cut -d':' -f1)
grep -q '==== Attachments ====' $DOKUWIKI_NAMESPACE/$FILE.txt sed -i $NUM,\$s/\[\[$TRAC_BS\/timeline?from=.*]] ago.//g $DOKUWIKI_NAME SPACE/$FILE.txt
# Formating attachments footer as Dokuwiki linkonly links
grep -q '==== Attachments ====' $DOKUWIKI_NAMESPACE/$FILE.txt sed -i $NUM,\$s/ * {{/{{/ $DOKUWIKI_NAMESPACE/$FILE.txt
grep -q '==== Attachments ====' $DOKUWIKI_NAMESPACE/$FILE.txt sed -i $NUM,\$s/|/?linkonly|/ $DOKUWIKI_NAMESPACE/$FILE.txt
grep -q '==== Attachments ====' $DOKUWIKI_NAMESPACE/$FILE.txt sed -i $NUM,\$s/]]/}}/ $DOKUWIKI_NAMESPACE/$FILE.txt
# In Dokuwiki, only lower characters are allowed in filenames
mv -u $DOKUWIKI_NAMESPACE/$FILE.txt `echo $DOKUWIKI_NAMESPACE/$FILE.txt| tr [:upper:] [:lower:]` 2/dev/null
done
tar -czvf $DOKUWIKI_NAMESPACE.tar.gz $DOKUWIKI_NAMESPACE
mkdir $DOKUWIKI_NAMESPACE-media
find $TRAC_PROJECT_ENV/attachments/wiki -type f -exec cp {} $DOKUWIKI_NAMESPACE-media \;
for FILE in `ls $DOKUWIKI_NAMESPACE-media`
do
# In Dokuwiki, only lower characters are allowed in media filenames
mv -u $DOKUWIKI_NAMESPACE-media/$FILE `echo $DOKUWIKI_NAMESPACE-media/$FILE| tr [:upper:] [:lower:]` 2/dev/null
done
tar -czvf $DOKUWIKI_NAMESPACE-media.tar.gz $DOKUWIKI_NAMESPACE-media
rm -rf $DOKUWIKI_NAMESPACE-media $DOKUWIKI_NAMESPACE
rm galletitas