[proposal] UDF to classify licenses

[vmarkovtsev]

This proposes to add a UDF based on go-license-detector to classify licenses of repositories to gitbase core. That UDF requires all the files in HEAD belonging to a repository.

cc @mcuadros @campoy

[campoy]

I'm not sure about this, what's the use case?

[vmarkovtsev]

@campoy The use case is to classify the licenses. Wdym?

[campoy]

I wouldn't want to add extra features that do not have a specific persona / use case in mind.
There's many useful things we could add to the engine, we need to ensure that those added have enough value to compensate for the extra API / documentation, etc.

[ajnavarro]

If that functionality requires all the HEAD files from a repository cannot be implemented using a UDF.

A UDF is working at the row level. You can process from 0 to N fields from a row, and generate a new value from that.

[vmarkovtsev]

It can work at the row level analyzing one file at a time @ajnavarro

[vmarkovtsev]

@campoy Regarding the functions. License audit is one of those nasty problems everybody hates to do in the enterprises yet many have to do (and pay for it).

[vmarkovtsev]

FWIW tagging @marnovo as he is supposed to decide on the product features.

[erizocosmico]

@ajnavarro what shall we do with this?

[ajnavarro]

If it is just implementing a library, we can create a function like LICENSE(repository_id)