yaml-to-bibtex.py is fragile

[rougier]

The yaml-to-bibtex script seems to be quite fragile and will break fro any typo in the yaml file (see for example #10). We need to make it more robust. Any taker ?

[schackartk]

I might like to take a stab at it (article.py as well). Are you using any code formatters or formal linters at this time?

I would like to start off by using a formatter (I would elect black, but happy to use whatever), and make sure it passes pylint and flake8. Then add type annotations and check with mypy.

I think those checks will get it pretty far in terms of robustness. From there, I would like to add more parsing checks and some unit tests (e.g. lastname and abbrvname extraction tests). Are there any objections to this proposal?

[rougier]

Thanks for the proposal. We're actually transitioning to a new system (JOSS) where metadata will be entered at time of submission. Hopefully we won't need the script anymore. But I retain your proposal in case things do not work as expected.

[schackartk]

Thanks! On that note, how soon do you think the transition will be complete? I have a replication I would like to submit soon and was wondering if it would be best to wait until after the transition?

[rougier]

Sorry for the delay. We have an early protype and I hope we'll make the transition by end of June but difficult to say. If you can wait a few more weeks, I should have a better timeline (don't hesitate to ping me in this thread).