[Proposal] A tiled FeatureGeometryLayer requesting and refining data based on zoom level

[airnez]

We lack of a way to display vector tile geometries depending zoom level. (except using rasterization with ColorLayers)
It would be nice to be able to refine geometries when zooming in. Here is the usual GIS data tiling schema for clarity :

image source

Context

Depending on zoom level, a vector tile service might simplify geometries like in this schema :

image source

Right now, FeatureGeometryLayers only request data once, at the layer min zoom level. If no min level is set, the data will be loaded with poor precision and will never refine. But setting a min level makes the layer disappear when dezooming, which is annoying when dealing with large-scale datasets.

This is mentionned in this issue : 2153

Description of the proposal

We could either:

• Modify the FeatureGeometryLayer API to enable / disable this behavior
• Implement a new layer with this behavior

[Desplandis]

It would be a nice feature to have LOD support for FeatureGeometryLayer. But we would have to aggregate same features (per id?) at different level within our LOD groups. We agree that geometry simplification will not be done on our side but from the dataset (for MVT at least)?