search

neuronets / trained-models

Trained TensorFlow models for 3D image processing
https://neuronets.dev/trained-models
22 stars 15 forks source link

resolve https://github.com/neuronets/trained-models/issues/87 #90

Closed hvgazula closed 1 year ago

gaiborjosue commented 1 year ago

@hvgazula See my latest PR

hvgazula commented 1 year ago

Please look into mine. It came earlier and is cleaner than yours. Thanks.

gaiborjosue commented 1 year ago

Hello @hvgazula the problem with your PR is that it does not consider the fact that the workflow can fail. If failed, the issue will be tagged as "failed" thus the user can't change the label manually to "Ready-to-test" since they are outside collaborators. Since you removed the file that does this automatically, there is no way of re-testing the workflow, it will just stay as "failed"

hvgazula commented 1 year ago

I am confused. Then what is this (see below)- You are asking the user to change the label. How is this possible? https://github.com/neuronets/trained-models/blob/b49f52df7c248295f6ad0205e4bcc568afbaae3d/.github/workflows/new_model.yml#L603-L607

gaiborjosue commented 1 year ago

Yes sorrry. I forgot to update the bot comment, however in the docs it does say to create an issue comment instead of changing the label.

hvgazula commented 1 year ago

Sure. Then what exactly is the assingIssue.yml doing? My understanding is as follows:

  1. Create issue using template ("ready-to-test" label gets added at issue creation)
  2. the issue gets assigned to the creator (this means the user/creator should be able to change labels). Am I right?
  3. If workflow fails, the label changes to "failed" (this is probably why the change_label.yml is needed)
  4. Once the user addresses the issues, they change the label to "ready-to-test" manually.
  5. workflow runs fine..label changes to "success"
gaiborjosue commented 1 year ago

No. They can't change labels manually. That is the purpose of change_label.yml