Backend API to store and retrieve zone polygons based on arbitrary geometry like administrative or survey data.
Start the web server with:
bundle exec rails server
The service is now available at http://localhost:3000 and the Swagger UI at http://localhost:3000/api-docs/index.html.
The only two API endpoints are:
- Fetch zone by id
- Fetch multiple zones selected by parameters
The returns are always geojson.
On all the endpoint the commons parameters are:
with_geom: include or not the geometries in the geojson.children_level: recursively include children of the zone and how much level.
The fetch multiple also have the following parameters:
bbox: restrict zones selection this bounding box. Affect the recessive children.intersect: restrict zones selection to ones intersecting the GeoJSON geometry. Default no limit.'property_filters: restrict zones selection the one having all this properties. Does not affect the recessive children.
curl -X GET "http://localhost:3000/v1/zones/5?with_geom=false" -H "accept: application/json"
{
"type": "Feature",
"properties": {
"id": 5,
"ancestor_id": 37666,
"name": "Marne",
"source": "OpenStreetMap",
"nature": "state_district",
"admin_level": 6
},
"geometry": null
}curl -X GET "http://localhost:3000/v1/zones/5" -d with_geom=false -d children_level=1 -H "accept: application/json"
{
"type": "FeatureCollection",
"features": [
{
"type": "Feature",
"properties": {
"id": 5,
"ancestor_id": 37666,
"name": "Marne",
"source": "OpenStreetMap",
"nature": "state_district",
"admin_level": 6
},
"geometry": null
},
{
"type": "Feature",
"properties": {
"id": 344,
"ancestor_id": 5,
"name": "Cormontreuil",
"source": "OpenStreetMap",
"nature": "city",
"admin_level": 8
},
"geometry": null
},
{
"type": "Feature",
"properties": {
"id": 345,
"ancestor_id": 5,
"name": "Reims",
"source": "OpenStreetMap",
"nature": "city",
"admin_level": 8
},
"geometry": null
},
...
]
}Using filter bbox=5.9718,49.8113,5.9718,49.8113 (in Luxembourg) reduced to one point and property_filters={"admin_level": 6}.
curl -X GET "http://localhost:3000/v1/zones" -d bbox=5.9718,49.8113,5.9718,49.8113 -d with_geom=false -d 'property_filters={"admin_level": 6}' -H "accept: application/json"
{
"type": "Feature",
"properties": {
"id": 89,
"ancestor_id": 181,
"name": "Canton Redange",
"source": "OpenStreetMap",
"nature": "state_district",
"admin_level": 6
},
"geometry": null
}Generate the OpenAPI spec
RAILS_ENV=test rake rswag:specs:swaggerize
Test the models and API controllers
bundle exec rails test
Test the OpenAPI Spec
bundle exec rspec
From the Zones API project directory.
Build Zones API Docker image:
docker-compose build
Build Cosmogony Docker
git clone https://github.com/osm-without-borders/cosmogony.git
cd cosmogony
git submodule update --init
docker build -t osmwithoutborders/cosmogony .
Build Cosmogony Explorer Docker importer
git clone https://github.com/osm-without-borders/cosmogony_explorer.git
cd cosmogony_explorer
docker-compose -f docker-compose.yml -f docker-compose.build.yml build importer
Initialize the Zones API database:
docker-compose up -d postgres && sleep 20
docker-compose run --rm web rake db:setup
Download an OpenStreetMap extract:
mkdir -p cosmogony_data
# Luxembourg +28 Mo
# France +4 Go
wget http://download.openstreetmap.fr/extracts/europe/luxembourg-latest.osm.pbf -O cosmogony_data/luxembourg-latest.osm.pbf
Compute zones hierarchy with Cosmogony from .osm.pbf, save the result as .json:
# Luxembourg 4 cpu, 40 s, +12 Mo
# France 4 cpu, 68 min, +390 Mo
docker run -v `pwd`/cosmogony_data:/data osmwithoutborders/cosmogony -i /data/luxembourg-latest.osm.pbf -o /data/luxembourg-latest.json
Import the .json into Postgres:
docker-compose -p cosmogony -f docker-compose.cosmogony.yaml up -d cosmogony_postgres
sleep 60 # Cool, Waiting for postgres to be ready
# Luxembourg 1 cpu, 11 s
# France 1 cpu, 14 min
docker-compose -p cosmogony -f docker-compose.cosmogony.yaml run --rm importer ./import.py import_data /data/luxembourg-latest.json
Export dump from Postgres, simplify using Lamber 93 projection (France, 2154):
docker-compose -p cosmogony -f docker-compose.cosmogony.yaml exec cosmogony_postgres psql -c "
COPY (
SELECT
id,
parent AS ancestor_id,
name AS name,
json_build_object(
'admin_level', admin_level,
'nature', zone_type
) AS properties,
'OpenStreetMap' AS source,
ST_Transform(
ST_SimplifyPreserveTopology(
ST_Transform(geometry, 2154),
ST_MaxDistance(ST_Transform(geometry, 2154), ST_Transform(geometry, 2154)) / 500
),
4326
) AS geometry,
now() AS created_at,
now() AS updated_at
FROM import.zones
) TO stdout" cosmogony cosmogony > cosmogony_data/luxembourg_cosmogony.tsv
Stop the Cosmogony part:
docker-compose -p cosmogony -f docker-compose.cosmogony.yaml stop cosmogony_postgres
Load the dump into the Zones API:
docker-compose run --rm web bash -c "cat ./cosmogony_data/luxembourg_cosmogony.tsv | psql -h postgres -c \"
COPY zones(id, ancestor_id, name, properties, source, geom, created_at, updated_at) FROM stdin;
SELECT setval('zones_id_seq', (SELECT max(id) FROM zones));
\" zone_api_development postgres"
Load adresses in Addok ndjson format. Download the French BAN adresses data base.
# + 500 MB
wget https://adresse.data.gouv.fr/data/ban/adresses/latest/addok/adresses-addok-france.ndjson.gz -P addok_data/
Import an adresses file.
docker-compose -f docker-compose.yaml -f docker-compose.addok.yaml run --rm import_addok rake ban_addresses:import[/addok_data/adresses-addok-france.ndjson.gz]
Link adresses to the city, so the cities can be the parent of the adresses.
docker-compose -f docker-compose.yaml -f docker-compose.addok.yaml run --rm import_addok rake ban_addresses:parenthood
Delete the imported addresses data.
docker-compose -f docker-compose.yaml -f docker-compose.addok.yaml run --rm import_addok rake ban_addresses:delete