geo-l1-capetown-building-centroids

Rerun

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely"]
# ///

import geopandas as gpd

def main():
    # Load the building footprints
    input_file = "capetown_buildings.shp"
    output_file = "building_centroids.geojson"
    
    gdf = gpd.read_file(input_file)
    
    # Check for building_id column. 
    # Often it might be 'id' or something similar. 
    # Let's check columns first.
    print(f"Columns in input: {gdf.columns.tolist()}")
    
    # We need to identify the ID column. 
    # Based on instructions, it's expected to be named 'building_id' or we should ensure it exists.
    # Let's see if 'building_id' is already there, or if we need to find the ID column.
    # The prompt says "ensure the output carries a building_id column with the building identifier from the input".
    # This implies there is an identifier in the input.
    
    # Common ID column names in shapefiles might be 'id', 'FID', 'OBJECTID', etc.
    # Let's assume the user meant the unique identifier of the building.
    # If 'building_id' is not present, we will look for common ID names or use the index.
    
    id_col = None
    possible_id_cols = ['building_id', 'id', 'FID', 'OBJECTID', 'ID']
    for col in possible_id_cols:
        if col in gdf.columns:
            id_col = col
            break
            
    if id_col is None:
        # If no obvious ID column, use the index
        print("No obvious ID column found. Using index as building_id.")
        gdf['building_id'] = gdf.index
    else:
        print(f"Using column '{id_col}' as the identifier.")
        gdf['building_id'] = gdf[id_col]

    # Ensure building_id is not empty (non-null)
    # Drop rows where building_id is NaN
    gdf = gdf.dropna(subset=['building_id'])

    # Calculate centroids
    # Note: centroid of a polygon might not always be inside it, but for buildings it's usually fine.
    # For better accuracy with complex geometries, we can use representative_point()
    # but 'centroid' is specifically requested.
    gdf_centroids = gdf.copy()
    gdf_centroids.geometry = gdf.geometry.centroid

    # Select only the required column: building_id and geometry
    # We keep only building_id to follow "ensure the output carries a building_id column" 
    # and "the addressing tool only renders points".
    # It's safer to keep only building_id and geometry to avoid clutter.
    output_gdf = gdf_centroids[['building_id', 'geometry']]

    # Write to GeoJSON
    output_gdf.to_file(output_file, driver='GeoJSON')
    print(f"Successfully wrote {len(output_gdf)} centroids to {output_file}")

if __name__ == "__main__":
    main()

# /// script
# requires-python = ">=3.12"
# dependencies = ["geopandas", "shapely"]
# ///

import geopandas as gpd

def main():
    # Load the building footprints
    input_file = "capetown_buildings.shp"
    output_file = "building_centroids.geojson"
    
    gdf = gpd.read_file(input_file)
    
    print(f"Columns in input: {gdf.columns.tolist()}")
    
    # The output showed 'building_i' which is likely a truncated 'building_id' 
    # due to Shapefile field name limitations (10 chars).
    
    # Let's look for anything that looks like an ID
    id_col = None
    possible_id_cols = ['building_id', 'building_i', 'id', 'FID', 'OBJECTID', 'ID']
    for col in possible_id_cols:
        if col in gdf.columns:
            id_col = col
            break
            
    if id_col is None:
        print("No obvious ID column found. Using index as building_id.")
        gdf['building_id'] = gdf.index
    else:
        print(f"Using column '{id_col}' as the identifier.")
        gdf['building_id'] = gdf[id_col]

    # Ensure building_id is not empty (non-null)
    # Drop rows where building_id is NaN
    gdf = gdf.dropna(subset=['building_id'])

    # Calculate centroids
    gdf_centroids = gdf.copy()
    gdf_centroids.geometry = gdf.geometry.centroid

    # Select only the required column: building_id and geometry
    output_gdf = gdf_centroids[['building_id', 'geometry']]

    # Write to GeoJSON
    output_gdf.to_file(output_file, driver='GeoJSON')
    print(f"Successfully wrote {len(output_gdf)} centroids to {output_file}")

if __name__ == "__main__":
    main()