latest version of lesson using "import imageio.v3 as iio" causes error

[shaw2thefloor]

In some examples, using this version of the imread command caused the following error:

maize = iio.imread("maize-root-cluster.jpg") maize[ maize < 128 ] = 0 fig, ax = plt.subplots() plt.imshow(maize)

---

ValueError Traceback (most recent call last) Cell In[18], line 1 ----> 1 maize[ maize < 128 ] = 0 2 fig, ax = plt.subplots() 3 plt.imshow(maize)

ValueError: assignment destination is read-only

Going back to the skimage.io.imread version fixes this issue.

[uschille]

Thanks for pointing this out @shaw2thefloor! The error can be avoided by adding maize = np.array(maize) after the line with iio.imread. It seems that I missed a few instances when implementing #247 - I will fix it as soon as I can. If you could point me to any other instances you found that would be a great help. Thank you!

[deppen8]

@uschille , a couple notes:

1. I went to try to make a fix here, but I couldn't find any examples where we weren't handling the error. Am I missing it somewhere?
2. The behavior/bug that leads to read-only images is changed in imageio as of v2.28.1 https://github.com/imageio/imageio/pull/976. The latest is v2.31.1, so if the learner creates an environment themselves, they will likely get a version of imageio that doesn't have the bug.
3. The version of imageio that ships with the base Anaconda distribution is different for Windows and MacOS 😱. https://docs.anaconda.com/free/anaconda/reference/packages/pkg-docs/. Anaconda on Windows still has imageio v2.26.0 (weird read behavior), but Mac and Linux users get imageio v2.31.1 (consistent read behavior).

I suppose, in light of point 3, we need to keep the np.array(img) work-around for now, but I put in a request to have Anaconda bump the version for Windows to match Mac+Linux (https://github.com/ContinuumIO/anaconda-issues/issues/13242). I have no idea what the constraints might be to doing that, so we shall see.

[mkcor]

@deppen8 thanks for taking care of this!

@shaw2thefloor the latest version of the lesson reads: """

# read input image
maize_roots = iio.imread(uri="data/maize-root-cluster.jpg")
maize_roots = np.array(maize_roots)

# display original image
fig, ax = plt.subplots()
plt.imshow(maize_roots)

Now we can threshold the image and display the result. PYTHON

# keep only high-intensity pixels
maize_roots[maize_roots < 128] = 0

""" Does it run properly for you?

[tobyhodges]

I am going to close this since we believe the issue had been resolved now. @shaw2thefloor please post back here if you think there is still an issue here, and we'll be happy to re-open.