I thought you might be interested that I created a multi-radical kanji lookup (component picker) from the KanjiVG data. It is available in the kanji learning app
https://shodoku.app/dictionary. To open the picker click the 部 button on the right hand side of the search bar.
It works exactly like the one on
jisho.org, except the data source is different, so not all radical lookups will yield the same kanji.
The index files are available to download (similar to kradfile and radkfile):
They are simple usv files like CSV that use unicode symbol values for separators (I have my reasons to not use CSV).
The code is kind of a mess at the moment, but I may rewrite the picker as a standalone web-component if anyone is interested in using it in their own project.
Note that
shodoku.app is a kanji learning app with a handy dictionary, so my main focus is not creating a dictionary. And the reason I decided to use KanjiVG as the data source for this picker (as opposed to radkfile and kradfile) was to stay consistent with the components displayed when studying each kanji. But in doing so I created a handy way to explore the graph of kanji-component mappings (and component-variation mapping) in the KanjiVG dataset.
I hope this picker and the kanji-component explorer can become useful for you all.