geo-l1-capetown-building-centroids

Rerun

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

def main():
    # Load the building polygons
    buildings = gpd.read_file("capetown_buildings.shp")
    
    print(f"Loaded {len(buildings)} buildings.")
    print(f"Columns: {buildings.columns.tolist()}")

    # Check for building identifier column. 
    # Based on prompt, I should look for something that can be assigned to 'building_id'
    # Since I don't know the exact name, I'll look for common ones or assume the user knows.
    # However, the prompt says "ensure the output carries a `building_id` column with the building identifier from the input."
    # I'll inspect the head of the dataframe first.
    
    # Let's find the ID column. Usually it's 'id', 'FID', 'OBJECTID', 'id_building', etc.
    # If I can't find a specific one, I'll have to guess or use the index.
    # But wait, I should probably just look at the columns.
    
    # Actually, I will try to find a column that looks like an ID or use the index if none found.
    # But standard practice is to check what's available.
    
    # Let's assume for now I need to identify the ID column.
    # I'll print columns and then decide. 
    # For the purpose of this script, I will check common names.
    
    id_col = None
    potential_id_cols = ['id', 'FID', 'OBJECTID', 'building_id', 'id_building', 'id_num', 'id_code']
    for col in potential_id_cols:
        if col in buildings.columns:
            id_col = col
            break
    
    if id_col is None:
        # If no obvious ID column, let's see what we have
        print("Could not find an obvious ID column. Using index.")
        buildings['building_id'] = buildings.index
    else:
        print(f"Found ID column: {id_col}")
        buildings['building_id'] = buildings[id_col]

    # Ensure building_id is non-empty/not null
    buildings = buildings.dropna(subset=['building_id'])

    # Create centroids
    # Note: centroid of a geometry might not be inside if it's very concave, 
    # but for buildings it's usually fine. 
    # Using representative_point() is safer to ensure the point is inside the polygon.
    centroids = buildings.copy()
    centroids.geometry = buildings.geometry.representative_point()
    
    # Keep only building_id and geometry
    centroids = centroids[['building_id', 'geometry']]

    # Save to geojson
    centroids.to_file("building_centroids.geojson", driver="GeoJSON")
    print(f"Successfully saved {len(centroids)} centroids to building_centroids.geojson")

if __name__ == "__main__":
    main()

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

def main():
    # Load the building polygons
    buildings = gpd.read_file("capetown_buildings.shp")
    
    print(f"Loaded {len(buildings)} buildings.")
    print(f"Columns: {buildings.columns.tolist()}")

    # Check for a column that might be the building ID.
    # 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.
    
    # First, check if there's a column that already looks like building_id.
    # If not, try to find another likely candidate.
    
    id_col = None
    # Priority list for potential ID columns
    potential_id_cols = ['building_id', 'id', 'FID', 'OBJECTID', 'id_building', 'id_num', 'id_code']
    
    for col in potential_id_cols:
        if col in buildings.columns:
            id_col = col
            break
    
    if id_col is not None:
        print(f"Found ID column: {id_col}")
        buildings['building_id'] = buildings[id_col]
    else:
        print("Could not find an obvious ID column. Defaulting to index.")
        buildings['building_id'] = buildings.index

    # Ensure building_id is non-empty/not null
    # Dropping rows where building_id is null as requested ("Every row must have a non-empty building_id")
    buildings = buildings[buildings['building_id'].notna()]
    
    # If it's a string column, empty strings might also be considered "empty"
    if buildings['building_id'].dtype == 'object':
        buildings = buildings[buildings['building_id'].astype(str).str.strip() != '']

    # Create centroids. 
    # We use the centroid method.
    centroids = buildings.copy()
    centroids.geometry = buildings.geometry.centroid
    
    # Keep only building_id and geometry
    centroids_output = centroids[['building_id', 'geometry']]

    # Save to geojson
    centroids_output.to_file("building_centroids.geojson", driver="GeoJSON")
    print(f"Successfully saved {len(centroids_output)} centroids to building_centroids.geojson")

if __name__ == "__main__":
    main()

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

def main():
    # Load the building polygons
    buildings = gpd.read_file("capetown_buildings.shp")
    
    print(f"Loaded {len(buildings)} buildings.")
    print(f"Columns: {buildings.columns.tolist()}")

    # The previous run showed 'building_i' as the column name.
    # Shapefiles have a 10-character limit for column names.
    
    id_col = None
    # Priority list for potential ID columns
    potential_id_cols = ['building_i', 'building_id', 'id', 'FID', 'OBJECTID', 'id_building', 'id_num', 'id_code']
    
    for col in potential_id_cols:
        if col in buildings.columns:
            id_col = col
            break
    
    if id_col is not None:
        print(f"Found ID column: {id_col}")
        buildings['building_id'] = buildings[id_col]
    else:
        print("Could not find an obvious ID column. Defaulting to index.")
        buildings['building_id'] = buildings.index

    # Ensure building_id is non-empty/not null
    buildings = buildings[buildings['building_id'].notna()]
    
    # If it's a string column, empty strings might also be considered "empty"
    if buildings['building_id'].dtype == 'object':
        buildings = buildings[buildings['building_id'].astype(str).str.strip() != '']

    # Create centroids. 
    # We use the centroid method.
    centroids = buildings.copy()
    centroids.geometry = buildings.geometry.centroid
    
    # Keep only building_id and geometry
    centroids_output = centroids[['building_id', 'geometry']]

    # Save to geojson
    centroids_output.to_file("building_centroids.geojson", driver="GeoJSON")
    print(f"Successfully saved {len(centroids_output)} centroids to building_centroids.geojson")

if __name__ == "__main__":
    main()