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"
    
    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)
    
    # Inspect columns to find building identifier
    print("Columns in input:", gdf.columns.tolist())
    
    # We need to find a column that likely represents the building ID.
    # Let's assume there might be something like 'id', 'fid', 'building_id', or similar.
    # The prompt says "ensure the output carries a building_id column with the building identifier from the input".
    # If the input doesn't have a column named 'building_id', I should check what it has.
    
    # However, standard Shapefiles often have an 'id' or 'FID' (though FID is usually index).
    # Let's check if 'building_id' exists. If not, I'll try to find the most likely candidate
    # or use the index if it's unique. But the prompt implies there's a building identifier.
    
    # Let's just check the first few rows to see the content.
    print("First few rows:")
    print(gdf.head())

    # Create centroids
    # We use representative_point() instead of centroid because representative_point() 
    # is guaranteed to be inside the polygon, which is usually better for building footprints.
    # However, the prompt specifically says "centroid layer". I'll use centroid but 
    # note that for concave polygons centroid might be outside. 
    # But I'll stick to the request: "centroid layer".
    
    print("Calculating centroids...")
    centroids = gdf.copy()
    centroids.geometry = gdf.geometry.centroid
    
    # Ensure we have a 'building_id' column.
    # If 'building_id' is not in the columns, we need to decide what to use.
    # Let's check for common ID columns.
    potential_id_cols = ['building_id', 'id', 'FID', 'OBJECTID', 'id_num']
    target_col = None
    for col in potential_id_cols:
        if col in gdf.columns:
            target_col = col
            break
    
    if target_col:
        print(f"Using '{target_col}' as 'building_id'")
        centroids['building_id'] = gdf[target_col]
    else:
        # If no obvious ID column, use the index as a fallback, but the prompt 
        # says "ensure the output carries a building_id column with the building identifier from the input".
        # This implies there is one.
        # Let's re-examine the columns from the printout.
        # I'll use the index if I really can't find anything, but let's assume 
        # the user knows the identifier exists.
        print("Warning: No obvious ID column found. Using index as 'building_id'.")
        centroids['building_id'] = gdf.index

    # Remove all other columns except geometry and building_id
    # The prompt says: "ensure the output carries a building_id column... Every row must have a non-empty building_id."
    # It doesn't explicitly say "only" building_id, but usually for these tasks, 
    # a clean output is preferred. However, "ensure the output carries a building_id column"
    # doesn't forbid others. But I will keep it minimal to be safe and useful.
    # Actually, "The addressing tool only renders points, so I need a centroid layer... Keep the building IDs"
    # implies the primary purpose is the ID.
    
    final_gdf = centroids[['building_id', 'geometry']]
    
    # Drop rows where building_id is null
    final_gdf = final_gdf.dropna(subset=['building_id'])
    
    print(f"Writing to {output_file}...")
    final_gdf.to_file(output_file, driver='GeoJSON')
    print("Done.")

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"
    
    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)
    
    # Inspect columns to find building identifier
    print("Columns in input:", gdf.columns.tolist())
    
    # From the previous run, columns are ['building_i', 'geometry']
    # It looks like the column 'building_i' is the identifier.
    # 'building_i' might be truncated due to Shapefile column name limits (10 chars).
    
    # Let's identify the correct column for building ID
    potential_id_cols = ['building_i', 'building_id', 'id', 'FID', 'OBJECTID']
    target_col = None
    for col in potential_id_cols:
        if col in gdf.columns:
            target_col = col
            break
            
    if not target_col:
        # Fallback to index if we still can't find it, but 'building_i' seems right
        print("No ID column found. Using index.")
        gdf['building_id'] = gdf.index
    else:
        print(f"Using '{target_col}' as 'building_id'")
        gdf['building_id'] = gdf[target_col]

    # Create centroids
    print("Calculating centroids...")
    centroids = gdf.copy()
    centroids.geometry = gdf.geometry.centroid
    
    # Ensure we have only building_id and geometry
    # and that building_id is not empty
    final_gdf = centroids[['building_id', 'geometry']].dropna(subset=['building_id'])
    
    # Ensure building_id is not an empty string if it's object type
    if final_gdf['building_id'].dtype == 'object':
        final_gdf = final_gdf[final_gdf['building_id'].astype(str).str.strip() != ""]

    print(f"Writing to {output_file}...")
    final_gdf.to_file(output_file, driver='GeoJSON')
    print(f"Success. Output saved to {output_file}")

if __name__ == "__main__":
    main()