ncells gets same value for all features in one timestep

[JuliaKukulies]

Thanks to @laurapaccini who posted in #245 that ncells in the feature output from the segmentation do not always correspond to the number of grid cells associated with the feature in the segmentation mask.

Eventually, I found the problem in the segmentation code, where we iterate over the different features in one time step, but assign only the value of ncells of the last feature to all features :

tobac/tobac/segmentation.py

Lines 267 to 275 in 09b176b

	# count number of grid cells asoociated to each tracked cell and write that into DataFrame:
	values, count = np.unique(segmentation_mask, return_counts=True)
	counts = dict(zip(values, count))
	ncells = np.zeros(len(features_out))
	for i, (index, row) in enumerate(features_out.iterrows()):
	if row["feature"] in counts.keys():
	ncells = counts[row["feature"]]
	features_out["ncells"] = ncells

I double-checked this behavior in our example notebooks, which all show only one unique value of ncells per timestep, although there are multiple features detected in each timestep.

This bug will be relatively easy to fix but should be done asap and preferably with a test making sure that this does not happen again.

Since we are about to release v1.5 maybe it suffices to include this in v1.5? But I leave it to @w-k-jones and @freemansw1 to judge if another hotfix is needed.

[freemansw1]

I'm inclined to release this as a 1.4.2 hotfix; I'd rather have a 1.4.x version that we know works available as we introduce the variety of new features in 1.5.0.

[w-k-jones]

I agree, happy to expedite this as a 1.4.2 release prior to 1.5

[JuliaKukulies]

True, that is a good point. I can take care of this

[JuliaKukulies]

This has been solved with #247 and will be released in v1.4.2