cthoyt / autoreviewer

👁️‍🗨️ Scientists often do the same bad stuff. Automate giving feedback during peer review.
MIT License
23 stars 0 forks source link

I like! #1

Open kjappelbaum opened 2 years ago

kjappelbaum commented 2 years ago

I feel it could be similar to checkcif, i.e. highlight potential issues for an expert reviewer (and required by journals).

I think one could check off many of the items on the JOSS reviewer checklist automatically.

Additionally, one might try:

As you suggested, one might want to also add subject-specific checks. For instance, for ML

Quite domain specific would then be to check if the train/test set are indeed independent.

kjappelbaum commented 2 years ago

Here's a tool that might already address some of the checkmarks https://github.com/fair-software/howfairis

cthoyt commented 2 years ago

I saw that. It doesn’t do what I want, and I don’t like how it does it. With this kind of stuff I would much rather rebuild it from scratch.