@Lauraschwarz brought to my attention that renaming keypoints/individuals could be useful in some workflows. For example they could load a dataset from file, rename things, and save it back to file.

This is possible with existing xarray functionality, but the syntax is not entirely obvious to those not familiar with xarray:

keypoint_names = ds.coords["keypoints"].values

# Create new keypoints with "snout" renamed to "nose"
new_keypoint_names = ["nose" if kp == "snout" else kp for kp in keypoint_names]

# Assign the modified values back to the Dataset
ds = ds.assign_coords(keypoints=new_keypoint_names)

We could document this workflow in one of our examples.

Alternatively, we could write convenience utilities (though these would be very thin wrappers), e.g.:

rename_keypoints(ds, {"snout": "nose", "tailbase": "tail_base"})
rename_individuals(ds, {"id1": "Alice", "id2": "Bob"})