I started using it last year and I think others would benefit from knowing about it too.

Offline dictionaries(based on data from wikipedia, wiktionary etc.):
https://github.com/itkach/slob/wiki/Dictionaries

I use the Wordnet and SimpleWikipedia dictionaries along with the one for the Wiktionary in my native language.


Going on a tangent about the dictionaries(you can skip this if you’re currently not interested on that):

They had a Malayalam dictionary, which was quite awesome for me, as I’m a Malayali. The dictionary is based on ml.wiktionary.org from 2021. Maybe there were not too many additions for the effort to make a new one.

The code for creating the dictionaries are also available in their github. I tried it out, but I’m on Opensuse Leap currently and installing couchdb seems tricky.

I also saw a python package, Pyglossary, which helps to convert dictionaries between opensource formats.

Wikipedia provides monthly dumps of zim files(used in Kiwix, another dictionary app. Not available in F-droid).
https://dumps.wikimedia.org/other/kiwix/zim/
I was able to use Pyglossary to convert those into slob files and use them.
So might be an easier way to get slob files, if slob files of Wiktionary or Wikipedia in your language are not available. Or if you want to update them. Data scraping is also avoided.

  • Achyu@lemmy.sdf.orgOP
    link
    fedilink
    arrow-up
    1
    ·
    5 months ago

    The fact that it is online

    Offline, right?

    It is very cool.

    The english wiktionary file is quite large. Are there other smaller versions?
    Wordnet is pretty good, so asking just to know if there are other dictionaries.

    • Blastboom Strice
      link
      fedilink
      arrow-up
      2
      ·
      edit-2
      5 months ago

      Offline, right?

      As long as you set Load remote content to never😅 (I think when this is turned on it loads stuff like images and such from wikipedia etc. which take unecessary cache space, so I leave it off).

      About the english wiktionary, oi… it’s huge (for comparison, the greek one is ~200mb).

      I don’t know if there’s a less complex version available anywhere… Maybe try creating an issue on aard2 github page. They seem to be active in general (btw, they also run an ftp server where they store ~all the dictionaries you see on github of multiple verions.) Also take a look here on the other projects, maybe you’ll find something useful (I’m taking a look at it now).

      As an extra topic, have you managed to convert wordnet dictionaries to .slob from pyglossary? I have found some ancient tools on github, but never made it (possibly you need to code some stuff).