Configure a region

Start from the template#

Copy gis/regions/TEMPLATE.yml to your own file and set GIS_REGION_CONFIG in .env. Define the region label, source ID, extent and layers. Check each source’s licence, attribution and permitted use independently.

Keep region-specific coordinates, agency names and identifiers in the region file and .env.

Supported source shapes#

A layer can use an ArcGIS item ID, a WFS 2.0 URL and type name, or an Overpass query. Optional source filters belong in the layer definition. The Overpass query defines its own area; the region extent is not applied to it.

Use attribute_hints to resolve varying source fields through the loader’s pick() helper. Do not add hardcoded source column names to general loader logic.

Inspect a small load#

docker exec gis-loader python gis/discover_layers.py --search reservoir
docker exec gis-loader python gis/load_gis.py --layer water_main --limit 500

The layer name must match your configuration. Inspect loaded raw attributes and conformed geometry before a full load.

Build topology and elevation#

docker exec gis-loader python gis/load_gis.py --all --build-topology
docker exec gis-loader python gis/load_elevation.py

Topology snaps pipe endpoints to shared nodes using PostGIS. Existing nodes retain IDs across reruns. Changing snap tolerance needs a deliberate topology rebuild and assessment of downstream references.

The DEM loader samples configured elevation tiles onto nodes without an elevation. Without a usable sample the model uses a default ground level; that is an assumption, not a measured elevation.

Supply and model selection#

The water model starts from recognised reservoirs and keeps connected supplied pipes. A mapped dam or tank is not automatically a valid fixed-head supply. Review facility classification and reservoir-to-node distance.

No valid reservoir or no supplied pipe causes model building to stop with a diagnostic. See Modelling before seeding and building a full network.