Example Dataset - Cropped FOV 0

[srivarra]

If you haven't already, please read through our contributing guidelines before opening your PR

What is the purpose of this PR?

Closes #817. Converts FOV 0 from a $1024 \times 1024$ image to $512 \times 512$.

How did you implement your changes

• Added the smaller FOV, reran through all notebooks, zipped up the necessary data files in the HF PR.
• Added a couple of instances of check_contrast=False in skimage.io.imsave.
• added a constant in settings.py for EXAMPLE_DATASET_REVISION, defaults to "main", change it to the commit ID to test a different branch.

Remaining issues

Will merge in after HF PR 7, HF PR 9 get merged.

• Merge HuggingFace PR 8

Need to address the pkl formats from Spatial LDA. pkl is unsafe and can allow for arbitrary code to execute when unpacked. Better to use a CSV or ZARR file for later.

[ngreenwald]

How similar were the individual pixie clusters? Was it pretty easy to match them up to the previous ones, or was there some tweaking needed?

[srivarra]

There was a little bit of tweaking, one of the clusters previously got merged into noise and something else. These are the results with the small FOV.

[ngreenwald]

kk, looks good