Use hvPlot in Plotting API, Add Polygon Projection Support

[philipc2]

Closes #849 #493 #918 #625

Overview

• Updates plotting API to use hvPlot
• Adds support for Cartopy Projections when plotting polygons (raster & vector)
• Updated to_geoataframe() methods to better align with recent changes to to_polycollection()
• Adds support for directly constructing a geopandas.GeoDataFrame using the engine='geopandas' parameter

[github-actions]

ASV Benchmarking

Benchmark Comparison Results

Benchmarks that have improved:

Change	Before [8b1c3de]	After [37abcf2]	Ratio	Benchmark (Parameter)
-	1.41±0.02μs	1.21±0.01μs	0.86	mpas_ocean.ConstructTreeStructures.time_ball_tree('120km')
-	969±20ns	870±10ns	0.9	mpas_ocean.ConstructTreeStructures.time_kd_tree('120km')

Benchmarks that have stayed the same:

Change	Before [8b1c3de]	After [37abcf2]	Ratio	Benchmark (Parameter)
	1.59±0s	1.61±0.01s	1.01	face_bounds.FaceBounds.time_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/mpas/QU/oQU480.231010.nc'))
	225±0.2ms	224±0.09ms	1.00	face_bounds.FaceBounds.time_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/scrip/outCSne8/outCSne8.nc'))
	2.03±0s	2.07±0s	1.02	face_bounds.FaceBounds.time_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/ugrid/geoflow-small/grid.nc'))
	7.57±0.07ms	7.65±0.04ms	1.01	face_bounds.FaceBounds.time_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/ugrid/quad-hexagon/grid.nc'))
	665±5μs	668±9μs	1.01	mpas_ocean.CheckNorm.time_check_norm('120km')
	433±6μs	448±8μs	1.03	mpas_ocean.CheckNorm.time_check_norm('480km')
	618±6ms	631±1ms	1.02	mpas_ocean.ConnectivityConstruction.time_face_face_connectivity('120km')
	40.9±0.3ms	41.6±0.9ms	1.02	mpas_ocean.ConnectivityConstruction.time_face_face_connectivity('480km')
	2.04±0.07ms	1.82±0.2ms	~0.89	mpas_ocean.ConnectivityConstruction.time_n_nodes_per_face('120km')
	495±10μs	498±20μs	1.01	mpas_ocean.ConnectivityConstruction.time_n_nodes_per_face('480km')
	342±30ns	315±1ns	0.92	mpas_ocean.ConstructTreeStructures.time_ball_tree('480km')
	321±20ns	299±10ns	0.93	mpas_ocean.ConstructTreeStructures.time_kd_tree('480km')
	1.06±0.01s	1.04±0.01s	0.98	mpas_ocean.GeoDataFrame.time_to_geodataframe('120km', False)
	57.9±0.3ms	54.8±0.7ms	0.95	mpas_ocean.GeoDataFrame.time_to_geodataframe('120km', True)
	78.5±2ms	77.5±0.6ms	0.99	mpas_ocean.GeoDataFrame.time_to_geodataframe('480km', False)
	5.28±0.1ms	5.08±0.08ms	0.96	mpas_ocean.GeoDataFrame.time_to_geodataframe('480km', True)
	2.69±0.01ms	2.69±0ms	1.00	mpas_ocean.Gradient.time_gradient('120km')
	288±0.6μs	289±2μs	1.00	mpas_ocean.Gradient.time_gradient('480km')
	248±4μs	245±8μs	0.99	mpas_ocean.HoleEdgeIndices.time_construct_hole_edge_indices('120km')
	123±1μs	124±1μs	1.01	mpas_ocean.HoleEdgeIndices.time_construct_hole_edge_indices('480km')
	407M	407M	1.00	mpas_ocean.Integrate.peakmem_integrate('480km')
	182±10ms	176±0.5ms	0.97	mpas_ocean.Integrate.time_integrate('120km')
	11.7±0.06ms	12.1±0.4ms	1.03	mpas_ocean.Integrate.time_integrate('480km')
	346±2ms	350±2ms	1.01	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('120km', 'exclude')
	346±2ms	346±2ms	1.00	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('120km', 'include')
	356±4ms	346±3ms	0.97	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('120km', 'split')
	21.2±0.1ms	23.0±0.3ms	1.08	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('480km', 'exclude')
	21.6±0.2ms	22.6±0.4ms	1.04	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('480km', 'include')
	21.6±0.1ms	22.6±0.4ms	1.05	mpas_ocean.MatplotlibConversion.time_dataarray_to_polycollection('480km', 'split')
	55.2±0.07ms	55.4±0.1ms	1.00	mpas_ocean.RemapDownsample.time_inverse_distance_weighted_remapping
	44.7±0.3ms	44.5±0.02ms	1.00	mpas_ocean.RemapDownsample.time_nearest_neighbor_remapping
	363±2ms	363±1ms	1.00	mpas_ocean.RemapUpsample.time_inverse_distance_weighted_remapping
	266±0.6ms	265±0.3ms	0.99	mpas_ocean.RemapUpsample.time_nearest_neighbor_remapping
	6.59±0.07ms	6.61±0.02ms	1.00	quad_hexagon.QuadHexagon.time_open_dataset
	5.66±0.02ms	5.69±0.06ms	1.01	quad_hexagon.QuadHexagon.time_open_grid

Benchmarks that have got worse:

Change	Before [8b1c3de]	After [37abcf2]	Ratio	Benchmark (Parameter)
+	380M	440M	1.16	face_bounds.FaceBounds.peakmem_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/mpas/QU/oQU480.231010.nc'))
+	381M	440M	1.15	face_bounds.FaceBounds.peakmem_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/scrip/outCSne8/outCSne8.nc'))
+	384M	442M	1.15	face_bounds.FaceBounds.peakmem_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/ugrid/geoflow-small/grid.nc'))
+	383M	440M	1.15	face_bounds.FaceBounds.peakmem_face_bounds(PosixPath('/home/runner/work/uxarray/uxarray/test/meshfiles/ugrid/quad-hexagon/grid.nc'))
+	1.69±0.01s	2.95±0.02s	1.74	import.Imports.timeraw_import_uxarray
+	263M	316M	1.2	mpas_ocean.Gradient.peakmem_gradient('120km')
+	240M	294M	1.22	mpas_ocean.Gradient.peakmem_gradient('480km')
+	369M	423M	1.15	mpas_ocean.Integrate.peakmem_integrate('120km')
+	236M	290M	1.23	quad_hexagon.QuadHexagon.peakmem_open_dataset
+	235M	289M	1.23	quad_hexagon.QuadHexagon.peakmem_open_grid

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[aaronzedwick]

Is there a better mesh you could use for the notebook example? Maybe with more cells? And possibly without the holes? It can be confusing to look at. Also now that I am thinking about it, we should start using the UXarray colormaps for the notebooks.

[philipc2]

Is there a better mesh you could use for the notebook example? Maybe with more cells? And possibly without the holes? It can be confusing to look at. Also now that I am thinking about it, we should start using the UXarray colormaps for the notebooks.

Agree with all the points here. I thought the PR was ready for review (but quickly turned it back to draft). I'll address this points!

[erogluorhan]

I've resolved the previous comments and now added a few more as part of my second round of review

[philipc2]

I've resolved the previous comments and now added a few more as part of my second round of review

Thank you!

[erogluorhan]

I am now good with all of the code here.

This PR seems to be reducing the code coverage though, and some of them looks real (e.g. _correct_central_longitude or part of it does not seem to be covered). I might be missing some contextvfrom the rest of the review process though. Could you clarify?

[philipc2]

I am now good with all of the code here.

This PR seems to be reducing the code coverage though, and some of them looks real (e.g. _correct_central_longitude or part of it does not seem to be covered). I might be missing some contextvfrom the rest of the review process though. Could you clarify?

Thanks for the comprehensive review! I've made a few changes since writing the initial tests as I've been addressing the reviews. I'll get those added and look into improving the code coverage.

[erogluorhan]

This is great work and marks a visualization milestone with the projections support it brings!