GeoSpace

A lightweight Python library for geospatial processing based on GDAL.

GeoSpace provides practical tools for raster and vector processing, including zonal statistics, reprojection, resampling, mosaicking, clipping, rasterization, and optional Google Earth Engine workflows.

Installation

uv add geospace

GeoSpace uses gdal-wheel as its GDAL dependency, so GDAL does not need to be compiled as part of GeoSpace.

import geospace
from osgeo import gdal

Example

import geospace as gs

rasters = ["precipitation.tif", "temperature.tif"]
basins = "basins.shp"

stats = gs.reduce(rasters, basins)
print(stats)

Features

• High-performance zonal statistics
• Raster reprojection, resampling, mosaicking, clipping, and nodata filling
• Raster and vector conversion
• Shapefile projection, filtering, buffering, and rasterization
• Coordinate and spatial calculations
• Optional Google Earth Engine integration

Optional dependencies

Google Earth Engine support:

uv add "geospace[gee]"

Mask-related features:

uv add "geospace[mask]"

Development

git clone https://github.com/xiejx5/geospace.git
cd geospace
uv sync

GeoSpace is a pure-Python package. GDAL wheels are built and published separately by gdal-wheel.