search

sparkfish / shabby-pages

ShabbyPages is a state-of-the-art corpus of born-digital document images with both ground truth and distorted versions appropriate for use in training models to reverse distortions and recover to original denoised documents.
MIT License
50 stars 6 forks source link

CI Failures - GitHub Releases #10

Closed proofconstruction closed 2 years ago

proofconstruction commented 2 years ago

Currently the GitHub Action for this repository is set to run daily, building a 50-image set of Shabby Pages and releasing this as a GitHub Release.

This is failing, because releasing this way requires tagging a commit, which we don't currently do. It's also not clear if we would need to commit to the repo every day, generating a new tag each time, and release against that, or if we could release against the same commit tag and still get a versioned history of files. Probably the former.

jboarman commented 2 years ago

Thanks @shotor. Your PR seems to have done the trick!

To recap, @shotor created a new workflow to run on a cron schedule that tags the lastest on dev with the current date. That tagging action then triggers the daily build workflow which creates a new release of 50 random shabby image patches.