Geo-semantic labelling of Open Data
$ git clone [email protected]:sebneu/geolabelling.git
$ cd geolabelling
- (optionally) setup virtual environment
$ virtualenv --system-site-packages geolabelling_env
$ . geolabelling_env/bin/activate
- Install requirements
$ python setup.py install
Download the GeoNames RDF dump and extract it to the "local" folder.
The GeoNames entities, labels, and additional/external links are all stored in a MongoDB instance. All commands provide --host and --port parameters to access the MongoDB
- Store all GeoNames entities and their parent relations in collection db.geonames:
$ python geonames_graph.py --host localhost --port 27017 geonames
- Build an index for all labels (and alternative labels) in collection db.keywords:
$ python geonames_graph.py --host localhost --port 27017 keywords
- Store country information in db.countries
$ python geonames_graph.py --host localhost --port 27017 countries
- Get postal codes from wikidata and store in collection db.postalcodes
$ python geonames_graph.py --host localhost --port 27017 wikidata-postalcodes
- Get NUTS from wikidata and store in collection db.nuts
$ python geonames_graph.py --host localhost --port 27017 wikidata-nuts
Adds the admin levels 2, 4 and 6 (according to the OSM hierarchy) to the DB. Uses the GeoNames API.
$ python geonames_graph.py divisions --country http://sws.geonames.org/3057568/
Adds the admin level 8 (districts, city divisions) to the DB
$ python geonames_graph.py city-divisions --country http://sws.geonames.org/3057568/
$ python openstreetmap/osm_inserter.py osm-polygons --level 8 --country http://sws.geonames.org/3057568/
$ python openstreetmap/osm_inserter.py poly-export --level 8 --country http://sws.geonames.org/3057568/ --directory poly-exports
First, download the OSM data for countries (e.g. from geofabrik.de) and place it in the "poly-exports/osm-export/" subdirectory. The script uses the polygons (e.g. poly-exports/slovakia/8) to extract streets and places for the subregions from the download, using Osmosis. The script takes as arguments the path to the repository and the admin level to extract the data. Also the path to Osmosis has to be modified in this script.
$ sh openstreetmap/osm_streets.sh /path/to/repo 8
Insert the data from the OSM extracts into the DB
$ python openstreetmap/osm_inserter.py insert-osm --country http://sws.geonames.org/3057568/ --level 8