rasterop 0.9.0

System to manage deforestation alert system

rasterop

Tiled raster processing utilities built on top of Rasterio and NumPy. This repository provides a small, extensible framework to apply custom operations on large raster datasets in a tiled/streaming fashion. It includes a generic tiled engine (tiled_op) and a few example operations you can adapt to your needs.

Features

• Process large rasters by windows (tiles) with multi-threaded reading/writing
• Compose custom operations by subclassing TiledOp
• Optional masking by geometry or raster mask
• Output tiling aligned to resolution ("target aligned pixels")
• Examples of common operations (median, thresholding, scaling, differences)

Requirements

• Python 3.12 or newer
• GDAL/GEOS system libraries required by rasterio and shapely (platform-specific)

Python package dependencies (runtime):

• rasterio
• numpy
• shapely
• pyproj
• tqdm
• scipy (used by TimeSeriesOp example)

Tip (Linux): You may need to install system packages like gdal, geos, and development headers before pip install rasterio shapely will succeed. See Rasterio/Shapely installation docs for your OS.

Installation

PyPI

pip install rasterop

Developement mode

Installation in development mode:

pip install -e .

License

See LICENSE.txt (MIT-style license). Copyright © 2025 CIAT.

Contributing

Contributions are welcome!

• Please open an issue to discuss substantial changes.
• Feel free to propose improvements to docs, tests, and examples.