Since meshes are large, it's not practical to serve them up via a simple http request similarly to routes.
The best solution I've found is probably to use an out of the box tile server, specifically maplibre/martin
Note: This is already under some exploratory development on the dw/serve_mesh_vector_tiles branch.
To do
Use postgis extension/container in docker compose stack,
Use martin container in docker compose stack,
On mesh save/create, create mesh polygon db entries, with cellbox metadata,
Consider managing cleanup of mesh polygons in db
Work out how to document in api schema,
In deployment config, enable postgis extension,
In deployment config, include running db script to add mesh vector tile (mvt) function
In deployment config, install additional dependencies for geodjango, e.g. PROJ, GDAL
In deployment config, set up martin as a service under systemd
In deployment config set up reverse proxy under apache/httpd
Document deployment
Integrate into frontend
Testing of django elements
Since meshes are large, it's not practical to serve them up via a simple http request similarly to routes.
The best solution I've found is probably to use an out of the box tile server, specifically maplibre/martin
Note: This is already under some exploratory development on the
dw/serve_mesh_vector_tilesbranch.To do