lazycogs is a Python package built by Development Seed for analyzing massive collections of cloud-optimized geotiffs. It builds on the pattern established by stackstac and odc-stac: take a STAC item collection and expose it as a spatially-aligned Xarray DataArray ready for parallel computation. stackstac and odc-stac are excellent tools that rely on the trusty combination of rasterio and GDAL for data i/o and warping operations and they cover many satellite imagery workflows well, but when you push them to work with thousands of assets they can struggle.

lazycogs takes the same approach but replaces GDAL and rasterio with python wrapped Rust libraries: rustac for STAC queries over stac-geoparquet files, async-geotiff for COG i/o, and obstore as the default cloud storage integration. The result is a tool that can instantly expose an Xarray DataArray view of massive STAC item archives in any CRS and resolution where the bytes are only fetched when you operate on the data for a given spatial/temporal window (i.e. lazy). Each array operation triggers a targeted spatial query on the stac-geoparquet file to find only the assets needed for that specific chunk - no upfront scan of every item required.

In this talk you will learn about why we built lazycogs, what it can be used for, and how the pattern of using narrowly scoped tools can be a useful alternative to systems built on monolithic projects.