Removing calls to skimage.viewer and switching to "lazy importing" module import - #99
Conversation
Sorry, something went wrong.
I have not removed/altered the lengthy section on skimage.viewer plugins, but it should probably be rewritten using something else (napari?) given that the viewer submodule will be deprecated on 0.18
…e snippets Section on Canny edge detection still has calls to skimage.viewer - it will need to be rewritten to stop relying on that submodule, since it will be deprecated in 0.18
tobyhodges
left a comment
There was a problem hiding this comment.
Thanks for submitting these changes @erickmartins. I agree that this is probably the best approach to take, though there are some further changes that will need to be made to the lesson afterwards, e.g. adding a callout box about interactive image viewers and including the output of calls to skimage.io.imshow in the lesson pages.
I only caught one change that still needs making, where the code block should be updated to match the adjusted explanatory text. See the suggestion included in my review. Once that is committed I will be ready to merge this.
Sorry, something went wrong.
Co-authored-by: Toby Hodges <tbyhdgs@gmail.com>
|
Should be done! Thanks. |
Sorry, something went wrong.
tobyhodges
left a comment
There was a problem hiding this comment.
Thanks @erickmartins. I will leave the PR unmerged for the next 24 hours, in case any of the other authors want to discuss this (fairly large) change further.
Sorry, something went wrong.
The Image Viewer submodule for scikit-image is being deprecated in 0.18 and removed in 0.20 (https://scikit-image.org/docs/dev/user_guide/viewer.html) - I've gone through the material (both episodes and example scripts) and changed all calls to skimage.viewer.ImageViewer with calls to skimage.io.imshow. Furthermore, any instances of import skimage have been replaced by specific imports for specific modules.
The Edge Detection episode has a lengthy section that uses the skimage.viewer Plugins feature - replacing that whole section felt outside the scope of this PR, so I've left it as is for the moment. It is probably a good idea to put a rewrite of that section using something else (napari?) in the roadmap.