opengeo 0.0.1

A Google Earth Engine-like Python package using Xarray and Dask

OpenGeo (og)

OpenGeo is a Python package designed to provide a Google Earth Engine-like experience using open-source geospatial tools such as Xarray, Dask, STAC, and GeoPandas. The goal is to smooth the transition for GEE users to the open Python ecosystem.

Features

• API Familiarity: Mimics ee.Image, ee.ImageCollection, ee.FeatureCollection and ee.Geometry APIs.
• Lazy Loading: Uses stackstac and dask to lazily load and process imagery from STAC catalogs.
• Local Processing: Run your analysis locally or on a standard Dask cluster without GEE quotas.
• Interoperability: Seamlessly integrate with xarray, geopandas, and shapely.

Installation

1. Create a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Install the package in editable mode:

   pip install -e .
   

Usage

import opengeo as og

# Initialize (optional, sets default STAC API)
og.Initialize()

# Create a region of interest
roi = og.Geometry.Rectangle(77.58, 12.96, 77.60, 12.98)

# Load an ImageCollection from a STAC ID (e.g., Sentinel-2 L2A)
col = og.ImageCollection("sentinel-2-l2a") \
    .filterDate("2023-01-01", "2023-01-31") \
    .filterBounds(roi) \
    .select(["red", "nir"])

# Compute a composite (mean)
image = col.mean()

# Calculate NDVI
ndvi = (image.select("nir") - image.select("red")) / (image.select("nir") + image.select("red"))

# Reduce over the region
stats = ndvi.reduceRegion(reducer="mean", geometry=roi, scale=10)
print(stats)

Comparisons

GEE	OpenGeo	Underlying Tech
ee.ImageCollection	og.ImageCollection	pystac_client + stackstac
ee.Image	og.Image	xarray.DataArray
ee.FeatureCollection	og.FeatureCollection	geopandas.GeoDataFrame
ee.Geometry	og.Geometry	shapely.geometry

Backend

OpenGeo defaults to using the Earth Search STAC API. You can change this via og.Initialize(url="...").

License

MIT