@urban-toolkit/autk-compute

---

@urban-toolkit/autk-compute / AutkComputeEngine

Class: AutkComputeEngine

Defined in: compute.ts:41

Main compute engine for GPGPU analysis and render sampling.

AutkComputeEngine provides the package-level API for running analytical and render compute passes. It accepts GeoJSON feature collections and writes the results back to feature.properties.compute on the returned collection.

Example

const compute = new AutkComputeEngine();
await compute.gpgpuPipeline(params);

Constructors

Constructor

new AutkComputeEngine(): AutkComputeEngine

Returns

AutkComputeEngine

Methods

gpgpuPipeline()

gpgpuPipeline(params): Promise<FeatureCollection<Geometry, GeoJsonProperties>>

Defined in: compute.ts:62

Runs the GPGPU pipeline and writes results to feature.properties.compute.

Parameters

params

GpgpuPipelineParams

Pipeline parameters.

Returns

Promise<FeatureCollection<Geometry, GeoJsonProperties>>

Promise resolving to the input collection with computed values attached.

Throws

If resultField or outputColumns is missing, or WGSL identifiers are invalid.

Example

const engine = new AutkComputeEngine();
const result = await engine.gpgpuPipeline({
  collection: fc,
  variableMapping: { pop: 'properties.population' },
  wgslBody: 'return pop * 1.5;',
  resultField: 'scaledPop',
});

---

renderPipeline()

renderPipeline(params): Promise<FeatureCollection<Geometry, GeoJsonProperties>>

Defined in: compute.ts:80

Runs the render pipeline and writes metrics to feature.properties.compute.render.

Parameters

params

RenderPipelineParams

Pipeline parameters.

Returns

Promise<FeatureCollection<Geometry, GeoJsonProperties>>

Promise resolving to the viewpoints collection with aggregated render metrics.

Throws

If no layers are provided or tileSize is invalid.

Example

const engine = new AutkComputeEngine();
const result = await engine.renderPipeline({
  layers: [{ id: 'b', collection: fc, type: 'buildings' }],
  viewpoints: { collection: vpFC },
  aggregation: { type: 'classes' },
});