predictor/examples/wind-demo/README.md

Wind layer demo

A minimal browser client that renders the predictor's wind field as an animated particle layer using Leaflet and leaflet-velocity.

The predictor's GET /api/v1/wind/field endpoint emits the wind-js-server "gfs.json" format (a two-element [U, V] array of {header, data} records), which is exactly what leaflet-velocity and sakitam-fdd/wind-layer consume — so no transformation is needed in the frontend.

Running

Serve this directory and the predictor from the same origin (or set API in index.html to the predictor's base URL and rely on the predictor's CORS headers):

# Terminal 1: the predictor (must have a dataset loaded for real data)
./bin/predictor

# Terminal 2: serve the demo
cd examples/wind-demo && python3 -m http.server 8090
# open http://localhost:8090  (set API="http://localhost:8080" in index.html)

API contract

GET /api/v1/wind/field query parameters (all optional):

Param	Default	Meaning
time	dataset epoch	RFC3339 forecast time to sample
altitude	0	altitude in metres
min_lat,max_lat,min_lng,max_lng	global	bounding box (degrees)
step	1.0	grid resolution in degrees (min 0.25)

GET /api/v1/wind/meta returns the active dataset's source, epoch, suggested altitudes, and bounding box so a client can populate its controls.

The full OpenAPI definition is served at /openapi.yaml, with a browsable ReDoc rendering at /docs.

Minimal fetch

const res = await fetch("/api/v1/wind/field?altitude=10000&step=2");
const data = await res.json();         // [ {header, data}, {header, data} ]
L.velocityLayer({ data }).addTo(map);