fishnet::geometry::OGRLayerGeometry Concept Reference

A geometry an OGR layer can be read into and written from. More...

#include <OGRGeometryAdapter.hpp>

Concept definition

template<typename G>
concept fishnet::geometry::OGRLayerGeometry = 
       std::same_as<G, Vec2DReal>
    || std::same_as<G, OGRRingAdapter>
    || std::same_as<G, OGRPolygonAdapter>
    || std::same_as<G, OGRMultiPolygonAdapter>
    || std::same_as<G, OGRGeometryAdapter>

Detailed Description

A geometry an OGR layer can be read into and written from.

The shapes are backed by the OGR geometry of the data source and copy nothing out of it. A point is the exception: it is two doubles, so Vec2DReal already is the cheapest representation there is and wrapping an OGRPoint behind a pointer would only cost an allocation.

OGRGeometryAdapter itself is included for layers which do not care what they hold: it keeps every geometry of the source, including the ones fishnet has no type for, and is what a pass through (merge, split, filter by attribute) should use. Its geometries carry no static GeometryType, so they have to be narrowed before any generic algorithm can be run on them.