FastArray 0.1.1

A lightweight Python package for preprocessing geospatial optical imagery for machine learning

fastarray

fastarray is a fast, lightweight Python package for preprocessing geospatial optical imagery (e.g., Sentinel-2) to prepare it for machine learning workflows. It simplifies operations like index computation (NDVI, NDWI, etc.), reprojection, patching, and transformation into PyTorch-ready tensors—while preserving geospatial coordinates.

---

🚀 Features

• ✅ Compute vegetation and water indices (NDVI, NDWI, etc.)
• ✅ Read raster data into coordinate-aware arrays
• ✅ Clip and reproject rasters using geometries (e.g., GeoJSON or Shapely)
• ✅ Split imagery into fixed-size patches
• ✅ Convert to PyTorch tensors for ML pipelines
• ✅ Fully Pythonic (no heavy GDAL/C++ dependencies)
• ✅ Fast and async-friendly where possible

---

📦 Installation

pip install fastarray