Improve Workflow for Tests and Transformation

[GusRiva]

We need to update the scripts as described here:

https://github.com/OpenStemmata/database/pull/98#issuecomment-1063875615

• I will update the transformation script so that it is triggered on pushes to the main branch and can also be triggered manually, and it just converts everything.
• We should unify the test scripts on pull requests in just one action. First the action should create the TEI files (without commiting them). Then the tests are done.

[Jean-Baptiste-Camps]

By the way, do we still need this branch : https://github.com/OpenStemmata/database/tree/transform ?

[GusRiva]

By the way, do we still need this branch : https://github.com/OpenStemmata/database/tree/transform ?

No, I just deleted it.

[GusRiva]

I'll try to get to work on this asap, while the thing is still fresh in my mind. Should I also try to "translate" your tests to python as was suggested? It's not absolutely necessary, as we can combine R and python in one run, but it would make sense to have the whole thing use only one language - and it's hopefully faster. I don't have any experience in creating that kind of tests with python but sounds like something interesting to learn.

[Jean-Baptiste-Camps]

Well, as you wish ! It certainly makes sense, and probably should be done in the end, but, on the other hand, we've done quite some engineering in the past days, and I would judge a priority to move forward on adding new stemmata. So, it's really a matter of what you feel like doing too.

[GusRiva]

Ok, I updated the tests in #109

I am leaving the R for now.

Currently the push to main is disabled, because it is a protected branch and we would need to create a secret access token. I did it in my test repo and it works. We would need to decide if you create the token, make me an admin so that I can do it or we create a user only for this purpose. We can discuss it next time we speak. For now, I would suggest merging #109 so that we can start contributing new stemmata with the tests working hopefully without generating errors. We can create the actual TEI files after we decide what to do with the token.

[Jean-Baptiste-Camps]

Ok, let's discuss it then ! Meanwhile, I made you co-owner of the organisation, which could facilitate things !

[Jean-Baptiste-Camps]

109 merged.

[GusRiva]

I got hooked into trying to do the tests in python and I think I did it 🥳

I tried them out in my clone repo and it seems to work really well. It takes about a minute to run, which is a great improvement! I'll open a pull request with the code

[GusRiva]

111

(Having your tests in R made the whole thing easier than I was anticipating and way easier than if I had started from scratch!)

[Jean-Baptiste-Camps]

Wahou, that was fast ! Congrats !