geo-l1-capetown-building-centroids

Rerun

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"Loaded {len(gdf)} buildings from {input_file}")
    print("Columns:", gdf.columns.tolist())

    # We need to find the correct building ID column. 
    # Based on the prompt, it should be named 'building_id' or we need to identify it.
    # Let's check the columns first. If 'building_id' is not there, we might need to find it.
    
    # The prompt says "ensure the output carries a building_id column with the building identifier from the input"
    # This implies the input might already have an ID column, or we might need to rename it.
    # Let's assume for now there's a column that represents the ID.
    
    # Let's check the first few rows to see what the identifier looks like.
    print(gdf.head())

    # Calculate centroids
    # Note: centroid calculation is more accurate if in a projected CRS. 
    # For Cape Town, EPSG:32734 (UTM 34S) is a good candidate if it's in WGS84.
    
    # We'll create a copy to avoid modifying the original if we want to keep original CRS for something else, 
    # but here we just want the centroids.
    
    centroids = gdf.copy()
    # Using representative_point() instead of centroid() is often better for polygons to ensure the point is inside.
    # However, the prompt asks for "centroid". Let's stick to centroid or representative_point.
    # "one centroid per footprint"
    centroids['geometry'] = gdf.geometry.centroid
    
    # Ensure we have 'building_id'. 
    # If the input has a different ID column, we should map it.
    # Let's check if 'building_id' exists.
    if 'building_id' not in centroids.columns:
        # If 'building_id' is not there, maybe it's 'id' or 'FID'?
        # Let's look for common ID columns.
        possible_id_cols = ['id', 'FID', 'OBJECTID', 'id_']
        found_id = False
        for col in possible_id_cols:
            if col in centroids.columns:
                centroids = centroids.rename(columns={col: 'building_id'})
                found_id = True
                print(f"Renamed {col} to building_id")
                break
        
        if not found_id:
            # If no obvious ID, use the index for now, but the prompt implies it exists.
            # Let's check the actual data first by printing columns.
            pass

    # Drop everything except 'building_id' and 'geometry' to keep it clean as requested
    # "The addressing tool only renders points... ensure the output carries a building_id column"
    # It doesn't explicitly say ONLY building_id, but it's a good practice for an addressing tool.
    # Let's keep only building_id and geometry.
    
    # First, make sure building_id is not null
    centroids = centroids.dropna(subset=['building_id'])
    
    # Keep only the necessary columns
    output_gdf = centroids[['building_id', 'geometry']]
    
    # Save 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()

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"Loaded {len(gdf)} buildings from {input_file}")
    print("Columns:", gdf.columns.tolist())
    print(gdf.head())

    # The prompt says: "ensure the output carries a building_id column with the building identifier from the input"
    # This suggests we need to identify which column is the identifier.
    # Looking at the print above will tell us.
    
    # Since I cannot see the output of the previous script until I run it, 
    # I will write a script that:
    # 1. Loads the data.
    # 2. Identifies the ID column.
    # 3. Calculates centroids.
    # 4. Creates the building_id column.
    # 5. Saves to GeoJSON.

    # Let's find the ID.
    # A common pattern in Shapefiles is 'id', 'fid', or specific names.
    # I'll search for anything that looks like an ID or just use the index if I must, 
    # but the prompt implies there is a building identifier in the input.
    
    # Let's try to find a likely candidate.
    potential_id_cols = [col for col in gdf.columns if col.lower() in ['id', 'fid', 'building_id', 'objectid', 'gid', 'uid']]
    
    if potential_id_cols:
        id_col = potential_id_cols[0]
        print(f"Using {id_col} as building_id")
    else:
        # Fallback: if no obvious ID, use the index
        print("No obvious ID column found. Using index as building_id.")
        gdf['building_id'] = gdf.index
        id_col = 'building_id'

    # Create centroids
    # Note: centroid can be outside the polygon for concave shapes. 
    # representative_point() is usually better for "one point per footprint" to ensure it's inside.
    # However, the prompt explicitly said "centroid". 
    # I will use centroid to follow instructions.
    
    centroids_gdf = gdf.copy()
    centroids_gdf['geometry'] = gdf.geometry.centroid
    
    # Map/Rename to building_id
    if id_col != 'building_id':
        centroids_gdf['building_id'] = gdf[id_col]
    
    # Filter out rows with empty building_id
    # If building_id is numeric, 0 might be a valid ID. 
    # If it's string, '' is empty.
    # If it's NaN, it's empty.
    centroids_gdf = centroids_gdf[centroids_gdf['building_id'].notna()]
    # Also check for empty strings if they are objects
    if centroids_gdf['building_id'].dtype == 'object':
        centroids_gdf = centroids_gdf[centroids_gdf['building_id'].astype(str).str.strip() != '']

    # Select only required columns
    output_gdf = centroids_gdf[['building_id', 'geometry']]
    
    # Save 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()

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"Loaded {len(gdf)} buildings from {input_file}")
    print("Columns:", gdf.columns.tolist())
    
    # The columns are ['building_i', 'geometry']
    # The identifier seems to be 'building_i'
    id_col = 'building_i'
    
    if id_col in gdf.columns:
        print(f"Found ID column: {id_col}")
    else:
        print(f"Warning: {id_col} not found. Checking for alternatives.")
        # Fallback if 'building_i' is not found (though the output showed it was)
        id_col = gdf.columns[0]

    # Create centroids
    # Using centroid as requested.
    centroids_gdf = gdf.copy()
    centroids_gdf['geometry'] = gdf.geometry.centroid
    
    # Set the building_id column
    centroids_gdf['building_id'] = gdf[id_col]
    
    # Remove any rows where building_id is null or empty
    centroids_gdf = centroids_gdf.dropna(subset=['building_id'])
    if centroids_gdf['building_id'].dtype == 'object':
        centroids_gdf = centroids_gdf[centroids_gdf['building_id'].astype(str).str.strip() != '']

    # Keep only the requested columns
    output_gdf = centroids_gdf[['building_id', 'geometry']]
    
    # Save 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()