Lars Grüter
Lars Grüter
Another point in favor of counter-clockwise being correct (made by @stefanv): ```python import skimage as ski tf = ski.transform.AffineTransform(rotation=np.deg2rad(90)) tf([1, 0]) # prints (x', y') # array([[6.123234e-17, 1.000000e+00]]) ```
Hey @vallsv, sorry on our part for letting this fall through the cracks! At a first glance this seems like it could be useful. I am not very familiar with...
I don't really have a feeling for how useful this might be. So I'd say it's up to you if you are interested in pursuing this. We can totally keep...
> The gallery is meant to show code the way you'd use it. I would make the distinction between two cases where we use randomness: 1. We use randomness to...
> But who gets notified of nightly failures, and how? The "Test Nightly" job is triggered on pushes to main. So the core dev creating the merge should be notified...
> Yes, I think that's the solution! Which one? The "raise an issue" one?
I don't see an issue with accessing `self.affines[i].params` in `__call__`. The attribute is part of the classes public API of `AffineTransform`. I agree that it doesn't seem as elegant, but...
> In any case, I'm happy to submit a PR if this change makes sense from an architectural perspective. @svepe if it's still convenient for you, please go ahead and...
Agreed! And sorry for the hiatus. Reviewer time can be a bottle neck. :/ Do you mind looking into the merge conflict? I can review this afterwards. :)
Oh and let us know if you have any questions.