Proposed new feature or change:
I found a tiny change we can make to speed up import uxarray. On my machine the import takes 1.7 seconds before this change, and 0.7 seconds after it. See also: #1260
The change is to import hvplot.pandas and hvplot.xarray lazily in uxarray.plot.accessor.py, i.e. only import them the first time any of the PlotAccessor objects' __init__ methods actually get called. These lazy imports should also make it easier to move hvplot to be an optional dependency instead of a required dependency later (see #1224).
This also delays when these strange gray bars appear in Jupyter (they are apparently caused by import hvplot.pandas and hvplot.xarray). Now they only appear the first time you plot something, instead of immediately upon importing uxarray for the first time.
Previous behavior:

(with first cell taking ~1.7s, second cell ~0.1s)
New behavior:

(first cell faster, second cell slower. Rerunning the second cell takes ~0.1s and does not produce the gray bars.)
Proposed new feature or change:
I found a tiny change we can make to speed up
import uxarray. On my machine the import takes 1.7 seconds before this change, and 0.7 seconds after it. See also: #1260The change is to import hvplot.pandas and hvplot.xarray lazily in
uxarray.plot.accessor.py, i.e. only import them the first time any of thePlotAccessorobjects'__init__methods actually get called. These lazy imports should also make it easier to move hvplot to be an optional dependency instead of a required dependency later (see #1224).This also delays when these strange gray bars appear in Jupyter (they are apparently caused by import hvplot.pandas and hvplot.xarray). Now they only appear the first time you plot something, instead of immediately upon importing uxarray for the first time.
Previous behavior:

(with first cell taking ~1.7s, second cell ~0.1s)
New behavior:

(first cell faster, second cell slower. Rerunning the second cell takes ~0.1s and does not produce the gray bars.)