⚠️ Use PgOSM Flex instead! ⚠️

This project is obsolete and unmaintained! Use the PgOSM Flex project instead

PgOSM: Making PostGIS + OpenStreetMap easier!

The goal of PgOSM is to make it simpler to load an OpenStreetMap .pbf file into PostGIS. The resulting data structure provides spatial analysts easier access to open-source spatial data. One focus of this project is to be easy to setup and run on low-power (e.g. non-server) hardware. The target machine for this process is a virtual machine within a consumer-grade laptop given 1GB ram and a single core processor.

See included LICENSE file for more details about licensing.

PgOSM via Docker

PgOSM is easiest deployed using the Docker image from Docker Hub. To use PgOSM without Docker, see the manual instructions (README-MANUAL.md).

Create folder for the output (~/pgosm-data), this stores the generated SQL file used to perform the PgOSM transformations and the output file from pg_dump containing the osm and pgosm schemas to load into a production database. The .osm.pbf file and associated md5are saved here. Custom templates, and custom OSM file inputs can be stored here.

mkdir ~/pgosm-data
mkdir ~/pgosm-input

To run custom transformations, place the SQL for the setup into ~/pgosm-input. The following command adds the included thematic_road transformation into the processing queue, thus into the SQL output at the end.

(Optional)

cp ~/git/pgosm/db/data/thematic_road.sql ~/pgosm-input/

To skip the default transformations, place a skip_default file into ~/pgosm-input.

(Optional)

touch ~/pgosm-input/skip_default

Start the pgosm container to make PostgreSQL/PostGIS available. This command exposes Postgres inside Docker on port 5433 and establishes links to local directories.

docker run --name pgosm -d \
    -v ~/pgosm-data:/app/output \
    -v ~/pgosm-input:/app/db/data/custom \
    -e POSTGRES_PASSWORD=mysecretpassword \
    -p 5433:5432 -d rustprooflabs/pgosm

Run the PgOSM Sub-region processing. Using the Washington D.C. sub-region is great for testing, it runs fast even on the smallest hardware.

docker exec -it \
    -e POSTGRES_PASSWORD=mysecretpassword -e POSTGRES_USER=postgres \
    -e PGOSM_SCHEMA=osm_dc \
    pgosm bash docker/run_pgosm_subregion.sh \
    north-america/us \
    district-of-columbia \
    400

Skip the PgOSM processing entirely with a skip_default file and no custom files. This is helpful for developing custom transformations.

Always download

To force the processing to remove existing files and re-download the latest PBF and MD5 files from Geofabrik, set the PGOSM_ALWAYS_DOWNLOAD env var when running the Docker container.

docker run --name pgosm -d \
    -v ~/pgosm-data:/app/output \
    -v ~/pgosm-input:/app/db/data/custom \
    -e POSTGRES_PASSWORD=mysecretpassword \
    -e PGOSM_ALWAYS_DOWNLOAD=1 \
    -p 5433:5432 -d rustprooflabs/pgosm

Why use PgOSM?

Before PgOSM

When data is loaded from osm2pgsql into your database you're left with a set of three (3) planet_osm_* tables split based on type of geometry, not by type of data.

public.planet_osm_point
public.planet_osm_line
public.planet_osm_polygon

This means that all the data representing waterways are found in the table public.planet_osm_line. So are all the roads, sidewalks, rivers, railways, and administrative boundaries. This setup makes it impossible for an analyst used to a properly normalized relational database to simply connect to the database and be productive.

After PgOSM

When pgosm runs it creates a schema named osm and loads layer-specific data into each table. The current default configuration generates 43 tables of logically grouped data. A few examples:

osm.road_line
osm.natural_point
osm.building_polygon
osm.waterway_line

The natural_point layer, for example, contains trees, peaks, and other items with nature-related tags in OpenStreetMap.

QGIS Layer Styles

See the QGIS styles README for details.