Better to use QVector here, because the QgsLineString/QgsGeometry
methods all use QVector too, and we want to avoid unnecessary
list->vector conversions.
These algorithms allow users to convert z or m values present in feature
geometries to attributes in the layer. By default the z/m value from the
first vertex in the feature is extracted, but optionally statistics
can be calculated on ALL the z/m values from the geometry (e.g. calculating
mean/min/max/sum/etc of z values).
Like the vector zonal stats algorithm, but this one works with
the zones defined in another raster.
Iterates over the input rasters in blocks to be nice and
memory efficient.
From the algorithm help:
"This algorithm calculates statistics for a raster layer's
values, categorized by zones defined in another raster layer.
If the reference layer parameter is set to "Input layer",
then zones are determined by sampling the zone raster layer
value at the centroid of each pixel from the source raster
layer.
If the reference layer parameter is set to "Zones layer",
then the input raster layer will be sampled at the centroid
of each pixel from the zones raster layer.
If either the source raster layer or the zone raster layer
value is NODATA for a pixel, that pixel's value will be
skipped and not including in the calculated statistics."
Adds a new interface QgsAbstractValidityCheck which defines
a single "check" which can be performed on a given QgsValidityCheckContext.
A new application-wide QgsValidityCheckRegistry registers
and manages instances of all known checks, and allows running
of all registered checks of a specific type at once.
Initially the framework is focused toward print layout validity
checks, but the interface has been designed to be generic enough
to allow alternative types of validity checks (e.g. project save
validity checks, processing model validity checks, etc.).
The API is designed to be used both by internal validity checks
and also to be extended by custom, organisation-specific
validity checks. E.g., for print layout validity checks we could have:
This is the processing side of the duplicate layer
names bug, while the core part was already fixed,
the processing logic was extended to handle inputs
from models and full-path references.
Fixes#20601
