worldgeo 1.3.2

world geohash index

worldgeo

World geohash index

A simple, easy to use python package to get a country by a given geohash or raw coordinates.

Getting started

Installation

pip install worldgeo

Usage

Import the Index class and load an index file (we'll discuss how to get one later on)

from worldgeo import Index
i = Index.load("world5.idx")

Now you can resolve a geohash

In [3]: i.find_by_hash("9smk")
Out[3]: 'MEX'

or coordinates

In [4]: i.find_by_coord(49.2827, -123.1207) # Vancouver
Out[4]: 'CAN'

Using a pre-built index

There are pre-built indexes based on https://github.com/johan/world.geo.json/blob/master/countries.geo.json . Given that, you don't really need to build indexes yourself, as there are ready to use files built by GitHub CI.

Pre-built indexes have various precision, check out the contents of prebuilt folder of this repo. Precision 6 means accuracy up to 1.2x1.2 km, which must be enough given the source might be even less accurate. Since more accurate index take much more disk and memory space, you might want to use precision 4 or 5 (39km or 5km correspondingly) as these take only a few Mb. Higher precision (7 and above) gives an index that is too large to fit in a default github repo, thus ShardedIndex was introduced. ShardedIndex splits the data into multiple files and can load them in parallel using multiprocessing.

Pre-built indexes can be loaded with a short-cut method load_prebuilt that takes an index name and a precision value. Available index names are world and vatsim. The world one resolves to a country alpha3 code and vatsim resolves to a region ICAO.

from worldgeo import Index

i = Index.load_prebuilt("vatsim", 5)
print(i.find_by_hash("ez656"))

The code above would resolve into LPPC-E which represents Lisboa FIR.

Building a new index

In case you still want to build a more (or less) precise index, you can do so by running

build-geoidx -o <output filename> -p <precision> [-t <num threads>] [-m] [-s]

and load the resulting file later on with Index.load(filename)