willyham/hashfill

Fill geofences with geohashes

Usage

Hashfill is a library for computing the set of geohashes which are contained by a geofence. It can either produce a set of hashes which are completely contained or one which also includes where the boundaries intersect. It currently operates on geom.Polygon objects from the https://github.com/twpayne/go-geom pacakge. Converting from geojson is simple and this package may offer a utility in the future.

func readFileAsGeometry(t *testing.T, file string) (*geom.Polygon, error) {
	data, err := ioutil.ReadFile(file)
	if err != nil {
		return nil, err
	}

	poly := new(geom.T)
	err = geojson.Unmarshal(data, poly)
	if err != nil {
		return nil, err
	}

	return nil, (*poly).(*geom.Polygon)
}

geofence := readFileAsGeometry("testdata/regents.geojson")
filler := hashfill.NewRecursiveFiller(
  hashfill.WithMaxPrecision(8),
)

hashes, err := filler.Fill(geofence, hashfill.FillIntersects)

Would result in something that could be visualized as:

Installation

This library depends on headers from the geos library. You can install with brew install geos or dnf install geos-devel on RPM based linux distros.

Options

• WithMaxPrecision - Sets the max hash precision the algorithm will generate hashes for.
• WithFixedPrecision - Causes the fully contained geohashes to still be divided into the hashes of the max precision.
• WithPredicates - Can be used to supply your own functions for Intersect and Contains rather than the built in ones.

Improvements

TODO:

• [] Fix bug in geofences with holes.
• [] Corner based check optimisations (+ hybrid)
• [] Cache poly to geom operation in predicates.
• [] Add benchmarks

hashfill's Issues

Go import error

While importing this package, getting the error.

github.com/paulsmith/gogeos/geos

In file included from ../../go/src/github.com/paulsmith/gogeos/geos/coordseq.go:4:
./geos.h:1:10: fatal error: 'geos_c.h' file not found
#include <geos_c.h>
^~~~~~~~~~
1 error generated.

And github.com/paulsmith/gogeos/geos page giving 404 error.

Is it it possible with Plus Codes?

Hello,
I want to know if the method of this library is possible with Open Location Codes?

All tiles returned with same precision.

Hi, as I understand your algorithm should return fully completed tiles with min precision and only uncompleted ones with max precision. May be there is misunderstanding from my side.
I always get array of geohashes with precision I defined as max precision.