geo-l1-tokyo-busstop-buffer

Rerun

import geopandas as gpd

def main():
    # Load the input file
    input_file = 'tokyo_connectors.geojson'
    output_file = 'tokyo_stop_catchments.geoparquet'

    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)

    # Ensure we are working with a projected CRS for accurate buffering in meters.
    # Tokyo is roughly around EPSG:3857 (Web Mercator) or a local UTM.
    # Let's check the current CRS first.
    print(f"Original CRS: {gdf.crs}")

    # If it's WGS84 (EPSG:4326), we must reproject to a metric CRS to buffer in meters.
    if gdf.crs is None or gdf.crs.is_geographic:
        print("Reprojecting to EPSG:3857 for buffering...")
        gdf_projected = gdf.to_crs(epsg=3857)
    else:
        gdf_projected = gdf.copy()

    # Create the 400m buffer
    print("Creating 400m buffer...")
    gdf_projected['geometry'] = gdf_projected.buffer(400)

    # Reproject back to the original CRS (usually WGS84) to maintain consistency
    if gdf.crs is not None and gdf.crs.is_geographic:
        print(f"Reprojecting back to {gdf.crs}...")
        gdf_buffered = gdf_projected.to_crs(gdf.crs)
    else:
        gdf_buffered = gdf_projected

    # The requirement is: "one Polygon or MultiPolygon row per input connector"
    # and "keep connector_id on every row". 
    # buffer() on a Point/Line/Polygon returns a single geometry per row.
    # However, if the input has multi-part geometries, buffer preserves them.
    # If we want to ensure we don't lose the connector_id, we just keep the columns.
    
    # Check if connector_id exists
    if 'connector_id' not in gdf_buffered.columns:
        print("Warning: 'connector_id' not found in input columns. Checking available columns...")
        print(f"Available columns: {gdf_buffered.columns.tolist()}")
        # If the user insists on connector_id, let's see if it's named something else or if they made a mistake.
        # For now, we'll assume it's there or the script will fail.

    # Save to GeoParquet
    print(f"Saving to {output_file}...")
    gdf_buffered.to_parquet(output_file)
    print("Done!")

if __name__ == "__main__":
    main()

import geopandas as gpd

def main():
    # Load the input file
    input_file = 'tokyo_connectors.geojson'
    output_file = 'tokyo_stop_catchments.geoparquet'

    print(f"Reading {input_file}...")
    gdf = gpd.read_file(input_file)

    # Ensure we are working with a projected CRS for accurate buffering in meters.
    # EPSG:3857 is a good general choice for meter-based operations globally, 
    # but for Tokyo, JGD2011 / Japan Plane Rectangular CS (e.g., EPSG:6677) might be more precise.
    # However, 3857 is standard for quick web-map tasks.
    
    # Let's stick to a metric projection.
    # For Tokyo, EPSG:6677 is highly accurate for the Tokyo area (Zone IX).
    # But for simplicity and robustness in this script, EPSG:3857 is fine.
    
    if gdf.crs is None or gdf.crs.is_geographic:
        print("Reprojecting to EPSG:3857 for buffering...")
        gdf_projected = gdf.to_crs(epsg=3857)
    else:
        gdf_projected = gdf.copy()

    # Create the 400m buffer
    print("Creating 400m buffer...")
    gdf_projected['geometry'] = gdf_projected.buffer(400)

    # Reproject back to the original CRS
    if gdf.crs is not None and gdf.crs.is_geographic:
        print(f"Reprojecting back to {gdf.crs}...")
        gdf_buffered = gdf_projected.to_crs(gdf.crs)
    else:
        gdf_buffered = gdf_projected

    # Verify connector_id exists
    if 'connector_id' not in gdf_buffered.columns:
        # If it's not there, let's see what's in the file. 
        # The user implies it should be there.
        print(f"Error: 'connector_id' column missing. Columns found: {gdf_buffered.columns.tolist()}")
        return

    # Save to GeoParquet
    print(f"Saving to {output_file}...")
    gdf_buffered.to_parquet(output_file)
    print("Done!")

if __name__ == "__main__":
    main()