versions.licenses in RubyGems Postgres Dumps

[zleonov]

Hi, not sure if Q&A is active here.

I have downloaded and imported the weekly dump of all gems from here: https://rubygems.org/pages/data.

I was surprised to see that the licensing data is not relational. I am not sure what format it's in at all.

Here is an example of the data in the licensing column:

---
- MIT
- GPL-2.0

Now, I can figure out that this gem is dual licensed under MIT and GPL2. Is this a standard format? If so what is it?

I see a lot of things that don't make any sense such as:

--- 
...

--- []

I am not worried about garbage data, but

1. I need clarification on how to parse this format, and
2. Am I suppose to assume this matches the response I would receive from queries the V1 or V2 API? (What I mean by this, is not whether the format will match, but whether the data will match e.g., whether the server will return [MIT, GPL-2.0] in the first example?

[simi]

It is yaml, it is free form and there is recommendation and also soft validation to use SPDX ID. See https://guides.rubygems.org/specification-reference/#license= for more details.

[zleonov]

Thanks @simi. I did not realize it was in yaml, but now that I look at it, I should have realized that.