pdok/sieve repository overview

⁠Sieve

Sieves GeoPackage⁠ Polygon geometries.

The purpose of this application is to prerefine the (MULTI)POLYGON geometries in a geopackage used for vector tiles by filtering out geometries (based on the given resolution) smaller than the pixels that are generated from the given vectoriles. By doing this specific artifacts/errors regarding the rendering of vector tiles can be omitted, and less data needs to be processed.

⁠Notes

• It will take a Geopackage and writes a new Geopackage where all the (MULTI)POLYGON tables are sieved.
  • All other spatial tables are 'untouched' and copied as-is.
  • Other not spatial tables are not copied to the new geopackage.
• The area of a POLYGON is used for determining if the geometries will be sieved, not the extent. So geometries with a extent larger then the given resolution but with an area smaller then that resolution will be sieved.
• A MULTIPOLYGON will be split into separate POLYGONs that will be sieved. So a MULTIPOLYGON containing elements smaller then the given resolution will have those parts removed.
• :warning: Spatialite lib is mandatory for running this application. This lib is needed for creating the RTree triggers on the spatial tables for updating/maintaining the RTree.
• It is also possible to replace the filtered out geometry with a centroid POINT geometry instead, using the -replace option

⁠Usage

go build .

go run . -s=[source GPKG] -t=[target GPKG] -r=[resolution for filtering] \
   -p=[pagesize for writing to target GPKG] -replace [t/f for replacing \
   instead of removing (optional)]

go test ./... -covermode=atomic

⁠Docker

docker build -t pdok/sieve .
docker run --rm --name sieve -v `pwd`/example:/example pdok/sieve ./sieve \
   -s=./example/example.gpkg -t=./example/example-processed.gpkg -r=50001 -p=10

With the docker example above processing the example.gpkg would result in the following.

⁠Inspiration

Code is inspired by the PostGis Sieve function⁠ from Mapbox.