larzgeo

Geospatial basics in pure Python. Zero dependencies.

The handful of geo operations most apps actually need - distance, bearing, geohashing, bounding boxes, and point-in-polygon - without GDAL, shapely, or any dependency.

from larzgeo import haversine, geohash_encode, point_in_polygon

haversine(51.5074, -0.1278, 48.8566, 2.3522)      # London->Paris ~343 km
geohash_encode(57.64911, 10.40744, 11)            # 'u4pruydqqvj'
point_in_polygon((2, 2), [(0, 0), (0, 4), (4, 4), (4, 0)])   # True

Why

• Just the essentials, correct. Great-circle (haversine) distance in km/mi/m/ nautical-mi, initial bearing, destination (project a point), midpoint, bounding_box (to pre-filter a radius query cheaply), geohash encode/decode (tested against the canonical u4pruydqqvj vector), and ray-casting point-in-polygon that handles concave shapes.
• No heavyweight geo stack. GDAL/shapely are powerful and huge; for "how far is X from Y", "is this point in this area", and "give me a geohash", this installs anywhere in seconds.
• Zero dependencies.

Install

pip install larzgeo

Usage

from larzgeo import (haversine, bearing, destination, midpoint,
                     bounding_box, geohash_encode, geohash_decode, point_in_polygon)

haversine(lat1, lon1, lat2, lon2, unit="km")     # or "mi" / "m" / "nmi"
bearing(lat1, lon1, lat2, lon2)                  # degrees, 0=N 90=E
destination(lat, lon, bearing_deg, distance)     # -> (lat, lon)
bounding_box(lat, lon, radius_km)                # (min_lat, min_lon, max_lat, max_lon)
geohash_encode(lat, lon, precision=12)
geohash_decode(hash)                             # -> (lat, lon)
point_in_polygon((x, y), [(x1, y1), ...])        # ray casting

Tests

python -m unittest discover -s tests -v   # 17 tests incl. geohash vector + concave polygons

The Larz stack

One of 30+ pure-Python, zero-dependency libraries at github.com/larz-scripter.

License

MIT (c) larz-scripter