Generating Voronoi polygons from polygon inputs (not points)

voronoi_polygons_on_polygon_data.py

import momepy
import geopandas as gpd

# Read data and convert to projected CRS
gdf = gpd.read_file(
    "/gdata1/projects/coastal/cem/sediment_compartments/compartments_checklist_final.geojson"
).to_crs("EPSG:3577")
col = "ID_Seconda"

# Create a limit for tessellation (join all polygons and buffer that)
limit = gdf.simplify(1000).buffer(50000).unary_union

# Run tesselation on specific column
# shrink: distance for negative buffer to generate space between adjacent polygons
# segment: maximum distance between points after discretization
tess = momepy.Tessellation(
    gdf, unique_id=col, limit=limit, segment=500, shrink=500
)

# Get the resulting GeoDataFrame
result = tess.tessellation

# Join tesselated data back into dataframe
joined_gdf = (
    gdf.drop("geometry", axis=1)
    .set_index(col)
    .join(result.set_index(col))
)
gdf_voronoi = gpd.GeoDataFrame(joined_gdf, geometry="geometry")

# Export to GeoJSON
gdf_voronoi.to_crs("EPSG:4326").to_file(
    "/gdata1/projects/coastal/cem/sediment_compartments/compartments_checklist_final_voronoi.geojson"
)

# Plot
ax = gdf_voronoi.reset_index().plot(column=col)
gdf.plot(ax=ax, column=col, alpha=0.5, linewidth=1, edgecolor="black")