Skip to content

Tutorials

Hands-on notebooks for learning Rasteret step by step.

Start with 01 and 02; they cover the most common workflows. Run rasteret datasets list to see all built-in datasets, then pick one for build(). 03 shows how to work with existing Parquet files that aren't from a STAC API. The rest cover filtering and custom configurations.

Tutorial What you'll learn
Quickstart: xarray build() a Collection, fetch pixels as xarray, compute NDVI
TorchGeo Integration Plug a Rasteret Collection into a TorchGeo training pipeline
Building from Parquet build_from_table() with real data from Source Cooperative + DuckDB exploration
Parquet-first Filtering Cache once, filter anywhere with Arrow predicates
Configuring Custom Collections Cloud configs, band mappings, and storage backends for non-built-in datasets
TorchGeo Benchmark Side-by-side performance comparison with native TorchGeo

Which build function?

Tutorials 01-02 use build(), which looks up STAC API details from the dataset catalog. Tutorial 03 uses build_from_table() for existing Parquet files. For custom STAC APIs not in the catalog, use build_from_stac(); see Collection Management.

Note

Some notebooks use live STAC queries and cloud reads. Re-execute locally with:

uv run python -m nbconvert --execute docs/tutorials/01_quickstart_xarray.ipynb