[JuhoErvasti]

Description

Mainly targeting GeneralizeBuildings. Main culprits were dissolve_and_inherit_attributes (and group_geometries_by_intersections_recursively) and reduce_nearby_points_by_selecting which are now optimized. The single largest remaining bottleneck is simplify_building_ruas which we can't directly affect (without writing our own implementation of it).

API changes:

• dissolve_and_inherit_attributes: remove inherit_from parameter, instead hard-code "most_intersection" option
  • "min_id" was kind of a legacy option and it was not used anywhere and I'm really hard-pressed to think why it would be ever used over most_intersection
• reduce_nearby_points_by_selecting:
  • remove priority_column, again this was not being used anywhere
  • force reference_points_gdf as a keyword argument; this is more in line with the library API generally
• group_geometries_by_intersections_recursively:
  • rename to group_intersecting_geometries
  • shorter and IMO clearer (and there's no longer a recursive function used)

Benchmark Change Report

Comparing perf-buildings against main (Threshold ±25.0)

Buildings:

Benchmark	Base Mean	PR Mean	Change	Status
test_generalize_buildings_50k	12.196 s	1.884 s	-84.55%
test_generalize_buildings_100k	4.528 s	536.22 ms	-88.16%

Others:

Benchmark	Base Mean	PR Mean	Change	Status
test_generalize_landcover	6.293 s	1.559 s	-75.22%
test_dissolve_polygons	11.69 ms	2.25 ms	-80.71%
test_generalize_building_areas	847.70 ms	533.28 ms	-37.09%
test_generalize_conservation_areas *	12.612 s	1.368 s	-89.16%

* This is misleading; the dissolve_and_inherit_attributes improvement probably improved things slightly, but I also reduced the amount of test data.

Type of change

• Bug fix

• New feature

• Other

• Non-breaking change

• Breaking change

Developer checklist

• A CHANGELOG entry has been included (only when change is visible to users)

[github-actions]

Benchmark Change Report

Comparing perf-buildings against main (threshold at ±25.0)

Passes threshold:

Benchmark	Base Mean	PR Mean	Change	Status
test_generalize_building_areas	1.202 s	809.70 ms	-32.65%
test_generalize_landcover	8.393 s	2.088 s	-75.12%
test_generalize_buildings_50k	16.813 s	3.324 s	-80.23%
test_dissolve_polygons	16.50 ms	3.23 ms	-80.43%
test_generalize_buildings_100k	5.814 s	790.08 ms	-86.41%
test_generalize_conservation_areas	17.098 s	2.126 s	-87.56%

Does not pass threshold

Benchmark	Base Mean	PR Mean	Change	Status
test_add_contiguous_lines_information[100]	1.31 ms	1.48 ms	+13.05%
test_merge_connecting_lines_by_attribute[5000]	68.46 ms	74.49 ms	+8.80%
test_merge_connecting_lines_by_attribute[1000]	16.10 ms	17.33 ms	+7.64%
test_count_connections[10000]	45.21 ms	47.50 ms	+5.08%
test_concatenate_lines	624.45 µs	655.56 µs	+4.98%
test_generalize_watercourse_lines	824.07 ms	853.82 ms	+3.61%
test_generalize_points[aggfunc]	14.48 ms	14.73 ms	+1.73%
test_generalize_shared_paths	3.103 s	3.142 s	+1.26%
test_assign_nearest_z[20000]	1.354 s	1.368 s	+1.06%
test_generalize_watercourse_areas	6.850 s	6.910 s	+0.88%
test_get_topological_points[100]	100.06 µs	100.67 µs	+0.61%
test_get_line_connection_points[100]	134.47 µs	135.12 µs	+0.48%
test_generalize_points[points]	14.19 ms	14.24 ms	+0.35%
test_generalize_polygons_to_points	1.77 ms	1.77 ms	+0.21%
test_generalize_slopelines	11.95 ms	11.97 ms	+0.13%
test_gaussian_smooth[10000000]	1.132 s	1.133 s	+0.05%
test_assign_nearest_z[5000]	315.05 ms	315.12 ms	+0.02%
test_generalize_contours	82.19 ms	82.18 ms	-0.00%
test_generalize_cliffs	14.96 ms	14.96 ms	-0.01%
test_assign_nearest_z[100]	6.47 ms	6.45 ms	-0.33%
test_generalize_points[displacement]	308.06 ms	306.84 ms	-0.40%
test_count_connections[100]	519.57 µs	517.37 µs	-0.42%
test_merge_connecting_lines_by_attribute[100]	4.15 ms	4.13 ms	-0.53%
test_get_contiguous_lengths[10000]	18.56 ms	18.46 ms	-0.56%
test_add_contiguous_lines_information[10000]	64.39 ms	63.95 ms	-0.68%
test_remove_overlap[points]	9.81 ms	9.74 ms	-0.72%
test_generalize_water_areas	1.288 s	1.278 s	-0.75%
test_merge_connecting_lines_by_attribute[10000]	147.22 ms	146.06 ms	-0.79%
test_flag_connections[100]	527.79 µs	523.49 µs	-0.82%
test_get_topological_points[5000]	3.58 ms	3.55 ms	-0.85%
test_generalize_railroads	1.462 s	1.449 s	-0.91%
test_gaussian_smooth[1000000]	80.38 ms	79.57 ms	-1.01%
test_substring	86.83 µs	85.92 µs	-1.05%
test_keep_intersection[points]	13.28 ms	13.14 ms	-1.05%
test_gaussian_smooth[100000]	8.05 ms	7.96 ms	-1.09%
test_get_line_connection_points[1000]	738.80 µs	729.93 µs	-1.20%
test_get_line_connection_points_roads	1.44 ms	1.42 ms	-1.22%
test_keep_intersection[lines]	13.66 ms	13.48 ms	-1.32%
test_get_contiguous_lengths[1000]	1.67 ms	1.65 ms	-1.37%
test_thin_points	64.47 ms	63.56 ms	-1.42%
test_keep_intersection[polygons]	15.31 ms	15.09 ms	-1.44%
test_assign_nearest_z[1000]	59.40 ms	58.49 ms	-1.53%
test_gaussian_smooth[30000]	2.43 ms	2.39 ms	-1.57%
test_remove_overlap[polygons]	11.83 ms	11.63 ms	-1.70%
test_flag_connections[1000]	3.12 ms	3.07 ms	-1.72%
test_get_contiguous_lengths[100]	390.23 µs	382.75 µs	-1.92%
test_count_connections[5000]	20.77 ms	20.37 ms	-1.92%
test_generalize_power_lines	3.539 s	3.471 s	-1.94%
test_remove_overlap[lines]	10.18 ms	9.98 ms	-1.99%
test_gaussian_smooth[1000]	258.81 µs	253.58 µs	-2.02%
test_get_topological_points[1000]	641.55 µs	627.48 µs	-2.19%
test_flag_connections[5000]	20.98 ms	20.51 ms	-2.22%
test_generalize_fences	753.73 ms	737.01 ms	-2.22%
test_get_contiguous_lengths[5000]	9.24 ms	9.03 ms	-2.31%
test_gaussian_smooth[100]	182.48 µs	178.16 µs	-2.37%
test_gaussian_smooth[10000]	928.52 µs	906.48 µs	-2.37%
test_get_line_connection_points[5000]	4.45 ms	4.34 ms	-2.52%
test_count_connections[1000]	3.04 ms	2.96 ms	-2.60%
test_get_line_connection_points[10000]	10.06 ms	9.75 ms	-3.08%
test_get_topological_points_roads	2.87 ms	2.78 ms	-3.20%
test_generalize_address_points	245.71 ms	234.76 ms	-4.46%
test_get_topological_points[10000]	9.17 ms	8.73 ms	-4.77%
test_assign_z_from_attribute	4.87 ms	4.55 ms	-6.58%
test_generalize_roads	12.41 ms	11.55 ms	-6.93%
test_add_contiguous_lines_information[5000]	33.61 ms	31.23 ms	-7.08%
test_add_contiguous_lines_information[1000]	5.80 ms	5.16 ms	-11.01%
test_flag_connections_to_reference	19.57 ms	17.35 ms	-11.33%
test_flag_connections[10000]	50.32 ms	43.62 ms	-13.32%