KerbSide

This documentation provides an overview of the KerbSide Generator, a tool that leverages the osm2streets Python wrapper to generate lane data including roads and sidewalks from OpenStreetMap (OSM). The tool allows users to download and process OSM tiles for a given location and outputs combined GeoDataFrames representing sidewalks, lanes, and intersections.

Features

• Retrieve bounding box and OSMID for a location using geopy.
• Download OSM tiles within a bounding box, split into smaller tiles.
• Process downloaded tiles to generate sidewalks, lanes, and intersections.
• Fix or remove problematic geometries in OSM data.
• Validate generated geometries for consistency and compatibility.
• Output results in GeoJSON format.

Example output:

Setup

Requirements

• Python 3.7+
• Install dependencies:

pip install geopy requests geopandas pandas

• Install osm2streets-python:

KerbSide currently works with a custom implementation of osm2streets-python which includes some upgraded error handling not currently present in the main project. Thus please install the package via the following command:

pip install --no-cache-dir -e git+https://github.com/lukasbeuster/osm2streets.git@python_wrap#egg=osm2streets_python\&subdirectory=osm2streets-py

Directory Structure

The script assumes the following directory structure:

• ../data/raw_data/osm2streets/location_cache.json: Stores cached location data.
• Output directories for tiles and processed files are automatically created under ../data/raw_data/osm2streets/{osmid}.

Usage

Running the Script

Use the command-line interface to specify the location, tile size, and driving side.

python KerbSide.py <location> [--tile_size <float>] [--driving_side <str>]

Arguments:

• location: Name of the location to process (e.g., "West, Amsterdam").
• --tile_size: Size of each tile in degrees (default: 0.01).
• --driving_side: Driving side of the region ("Right" or "Left", default: "Right").

Example:

python KerbSide.py "West, Amsterdam" --tile_size 0.02 --driving_side Left

% TODO: add example figures

Key Functions

1. Location Handling

get_location_info(geolocator, location_name)

Retrieves OSMID and bounding box for the specified location using the geopy geolocator.

load_location_cache() / save_location_cache(cache)

Handles caching of location data to avoid redundant API calls.

2. Tile Management

download_tiles(osmid, bbox, tile_size, output_dir)

Downloads OSM tiles within the specified bounding box, dividing the area into smaller tiles of the given size.

• Parameters:
  • osmid: OSM ID of the location.
  • bbox: Bounding box (min_lat, max_lat, min_lon, max_lon).
  • tile_size: Tile size in degrees.
  • output_dir: Directory for saving the tiles.

3. Data Processing

process_tiles(tile_dir, input_options, output_dir)

Processes the downloaded OSM tiles to generate GeoDataFrames for sidewalks, lanes, and intersections.

• Parameters:
  • tile_dir: Directory containing OSM tiles.
  • input_options: Configuration for osm2streets.
  • output_dir: Directory for saving processed GeoDataFrames.

fix_or_remove_invalid_ways(osm_file, output_file)

Identifies and fixes or removes problematic ways in OSM data.

validate_geometry(geojson_func)

Validates geometries generated by osm2streets and ensures they are compatible with GeoPandas.

4. Configuration Options

The script uses osm2streets with customizable input options:

• debug_each_step: Enable detailed debugging steps.
• dual_carriageway_experiment: Test dual carriageway handling.
• sidepath_zipping_experiment: Test sidepath zipping.
• inferred_sidewalks: Generate inferred sidewalks.
• inferred_kerbs: Generate inferred curbs.
• override_driving_side: Specify driving side ("Right" or "Left").

Output

Generated Files

The script outputs the following GeoJSON files:

1. combined_network.geojson: Combined sidewalk network.
2. combined_lanes.geojson: Lane polygons.
3. combined_intersections.geojson: Intersection markings.

Logs

•	failed_tiles.txt: List of tiles that could not be processed.

Error Handling

The script includes mechanisms to:

• Cache and reuse location data to minimize API queries.
• Detect and fix invalid geometries in OSM data.
• Skip invalid .osm files (empty, etc.) and log failures.

Known issues

• some osm files cause the

Contributing

Contributions are welcome. Just open issues / pull requests.