unwarped images in VBN session object stimulus_templates have missing (NaN) pixels

[corbennett]

Describe the bug The VBN session object stimulus_templates look fine for the warped images, but the unwarped images have NaNs where they shouldn't. Displaying these shows holes in the image that aren't there for the warped templates (and weren't on the display for the animal during experiments).

To Reproduce

cache = VisualBehaviorNeuropixelsProjectCache.from_s3_cache(
            cache_dir=cache_dir)

session_id = 1053941483
session = cache.get_ecephys_session(
            ecephys_session_id=session_id)

image = np.copy(session.stimulus_templates['unwarped']['im034_r'])
plt.imshow(image,cmap='gray')

This will produce the following image:

The white patches are NaNs in the array. The warped image doesn't have these holes:

Expected behavior I expect the unwarped image to look like the pre-warped warped image.

[morriscb]

Already fixed in current release 2.14.1 SDK release.