Dominik Moritz
Dominik Moritz
I think keeping the aliases is very valuable so I'd say option 1.
I think it's a trade off. Sometimes people generate schemas for code they do not control so relying on annotations for features of the schema may not be the right...
I implemented @hidden five years ago https://github.com/vega/ts-json-schema-generator/issues/329
Vega fusion already does something like this for Vega in python. Nonetheless, it would be cool to adopt this for js as well.
You could develop a loader similar to the Vega arrow loader as an external package. To you want to take a stab?
Start small. You could manually project columns in the loader as a start.
That's super cool. I'm hesitant to rely on the package in core for now but could be worth making an extension. The compressor package is more than 100kb, though: https://github.com/hyparam/hyparquet-compressors.
Yep. Should be easy. If you want to help set that up, that would be great. I was thinking of maybe bundling flechette with Vega in a coming version but...
> Broader question: would we be interested to have portions of the vega sub-packages converted to TS so we don't have to sync .d.ts files with JS source code as...
The schema url is not evaluated by Vega I thought. Can you explain what command throws an error? You might need a relative url like ./etc