osgende is a Python-framework for creating customized Postgresql databases from OSM data. It expects the OSM data to be already in a random-accesibly storage and extracts and updates relevant data from there. Currently, the only backing storage implemented is a Osmosis-like Postgresql database with a file-based node-location storage provided by libosmium.

Osgende requires Python3.

• PostgreSQL >= 9.5 https://www.postgresql.org

  Postgres database.

• psycopg >= 2.5.0 http://initd.org/psycopg/

  Python bindings for PostgreSQL.

• Shapely http://trac.gispython.org/lab/wiki/Shapely

  Python bindings for the geos library. (available as Debian package: python-shapely)

• pyosmium >= 3.0 https://github.com/osmcode/pyosmium

  Python bindings for libosmium, needed for the import tool.

• SQLAlchemy >= 1.4 http://www.sqlalchemy.org/

  SQL toolkit for python. osgende uses the Core package only.

• GeoAlchemy2 http://geoalchemy-2.readthedocs.org

  Postgis extension for SQLAlchemy.

Optionally, the following packages are required by some of the tools:

• Mapnik >= 3.1 http://mapnik.org/

  Map renderer for OSM data.

Osgende can be installed using Python's default setup tools:

pip install .

Osgende always needs a backing database that contains a full copy of the OSM data you like to process. You can use the osgende-import tool to create such a database. To create a new database and import an OSM file into the database run:

osgende-import -d planet -c -i liechtenstein.osm.pbf

It is strongly recommended that you make use of an external node location file to speed up processing using the -n option.

You need to create your own MapDB and instances of tables. For an example see the waymarked-trails project.