Will stack graphs definitions for python language be open sourced soon?

[hetulvp]

The stack-graphs/languages/ dir currently contains stack graphs definitions for below two languages.

1. tree-sitter-stack-graphs-java
2. tree-sitter-stack-graphs-typescript

Can you open source definitions for the python language as well?

cc @dcreager @hendrikvanantwerpen

[dcreager]

Yes, this is definitely something we want to do, though for some definition of "soon"...

The Python rules work, and are running in production, but use an older prototype precursor to the TSG language. We would have to port those rules over to the new DSL, and have not had the time to do so. We could put a copy of the old-DSL Python rules into this repo, though I'm not sure how useful that would be — there's nothing secret in them, but they might be inscrutable since there are subtle differences between the old DSL and the new TSG.

[hetulvp]

Thanks for the prompt response. I would like to contribute if you can add a copy of the old-DSL Python rules into this repo or a different branch for now and give me some pointers through basic conversion examples to start with. Would that be a good idea?

Thanks again.

[hendrikvanantwerpen]

@hetulvp Thanks for offering to contribute. I'll see what we have internally and how to best share trust with you. Then I'm willing to act as your support contact on this.

Just to set expectations, some amount of figuring things out yourself will probably be involved given my own time. But I'll try to help you as much as I can within my own time constraints.

Having a public Python definition here would be really cool, and I'm excited if we can make that happen!

[hendrikvanantwerpen]

@hetulvp I did some preparatory work to get a basic conversion into shape. The result is in #282. As you can see some of the tests are still failing, so it doesn't behave exactly the same as the legacy rules. But that should be good starting point to dive in if you want!

(The legacy rules are also included in stack-graphs.scm. Unfortunately, the legacy implementation is not something I can easily share to test it's behavior, but many of the concepts are broadly the same.)

[hetulvp]

Hey @hendrikvanantwerpen This is awesome !!! Thanks for taking your time out and sharing the legacy implementation. I will start looking into this now and let you know if I need any help understanding any part of this file.