Sunday, November 10, 2019

WikDict source code now at GitHub

From Mercurial to Git

Since BitBucket stopped support for Mercurial, I finally accepted that git has won the majority of developer mindshare. As a result of this, WikDict now uses git and the repositories have moved to GitHub:

Easier Dictionary Generation

The hardest part of generating the WikDict dictionaries on a new system was usually getting the installation of the Virtuoso database right. Different versions include different bugs and many Linux distributions don't contain packages for it or only outdated ones. But thanks to the excellent TenForce Virtuoso docker image, this can now be solved uniformly across all docker-capable systems. The changes in WikDict are small and should make the data loading as simple as running a few scripts.

Wednesday, February 27, 2019

WikDict now has more than 6 million translations!

The Wiktionary editors diligently added new translations, the dbnary project fixed bugs in the data extraction and I took advantage of all this to update the WikDict data set. The result is a higher quality set of translations with over 6 million different translation entries.
Thanks to Hannes Pirker's contribution to dbnary, we also have pronunciation information for German words for the first time.