SOLVED:
I got the stardict-tools (GUI and command line tools) from Arch user repos.
To remove HTML and make a nice looking, Duokan-compatible dictionary:
- Install stardict-tools-git from AUR.
- "Decompile" your trio of stardict files (whatever.dict, whatever.idx, whatever.ifo) to a "tabfile" using stardict-editor GUI. (Select the .ifo file.) This will give you a text file whatever.tab
- Use sed, python, whatever script to remove HTML crap from this file. Search around for a regex to strip HTML tags if you don't want to think about it. For example:
Code:
sed -e 's/<[^>]*>//g' whatever.tab
Afterward, you might want to search and remove/replace in a text editor nonstandard(?) stuff like ' and "
- Run
Code:
stardict-tabfile whatever.tab
to generate the new dict, idx, ifo files. You can open the ifo file in a text editor to customize the dictionary name. Leave the other lines alone.