DCbrain makes public some of its code in open source
DCbrain makes public some of its work, specifically a data converter, on its github account. It is a postgresql geographic data converter in openstreetmap format (which DCbrain uses in its Artificial Intelligence INeS solution) to power tools in this ecosystem.
The added value is in particular the automation of conversion, for faster and more mass processing. DCbrain uses it in particular to convert enriched road data to power OSRM software. This converter has the following name: pgsql2osm.
The code is intended for developers who use software from the OpenStreetMap free ecosystem (OSM).
It automates and makes it possible to extract large amounts of data from a Postgresql database to the XML OSM file format. It allows DCbrain today to power their OSRM (http://project-osrm.org/) routing platform with pre-processed data in their Postgresql database. Usually, OSRM is powered directly via OpenStreetMap while DCbrain prefers to use its own data for their use.
“The openStreetMap community, software ecosystem is little known to businesses and we want to create the link between these two worlds. Our approach is part of cooperation and is important to contribute to the community’s efforts. The latter allows us to have effective and free solutions, we must be able to contribute to them as well. We think above all of “ecosystem.” We are confident enough about the quality of our work, our technical expertise and know-how, to allow us to share it,” comments Benjamin de Buttet, COO and co-founder of DCbrain.
Obviously aware that it is a relevant data source, DCbrain uses OpenStreetMap (OSM) and its tools to produce route calculations, dashboards and map renderings. With this approach, DCbrain contributes to the common effort. By releasing the code from certain developments, including an implementation of a postgresql conversion to xml osm, licensed GNU GPL v3. pgsql2osm allows you to produce xml osm files from postgis geometries (linestrings and dots).
This implementation uses a few SQL queries to turn geometries into topological postgresql tables resembling the OSM model.
It manages the connections between the different paths with its own layers and the Python and then converts it all into XML OSM. DCbrain now uses it in production to work regionally with datasets of about 1M lines and 7M nodes on a 4CPU/16GB RAM server.
Have a look at our github here: https://github.com/DCbrainOrg/pgsql2osm