Python API reference

Import the functions below from pygeofilter_aeronet, except to_aeronet_api, which is in pygeofilter_aeronet.evaluator. Filters accept a CQL2 JSON string or a mapping.

Function	Result / side effects
`to_aeronet_api(cql2_filter)`	Returns `(query_string, query_parameters)` for the generated HTTP client
`dry_run_aeronet_search(cql2_filter, url=AERONET_API_BASE_URL)`	Logs the observation URL; returns `None`
`aeronet_search(cql2_filter, output_dir, url=AERONET_API_BASE_URL, verbose=False, timeout=None)`	Writes CSV and GeoParquet; returns a `pystac.Item` with `csv` and `geoparquet` assets
`get_aeronet_stations(url=AERONET_API_BASE_URL, verbose=False, timeout=None)`	Downloads station metadata; returns a list of `pystac.Item` objects
`dump_items(items, output_file)`	Writes STAC Items to GeoParquet; returns `None`
`query_stations_from_parquet(file_path, cql2_filter=None)`	Returns `(sql_query, items)`; no filter selects all stations

Pass pathlib.Path objects for output_dir and output_file, and a string path or URL for file_path. Python HTTP helpers default to no timeout; the CLI defaults to 30 seconds.

AERONET_API_BASE_URL is https://aeronet.gsfc.nasa.gov. DEFAULT_STATIONS_PARQUET_URL points to https://github.com/Terradue/pygeofilter-aeronet/raw/refs/heads/stations-update/stations.parquet.

Importing the package installs/loads DuckDB's spatial extension and reads the hosted station catalogue to validate site names. Even query translation and dry runs therefore require those resources.

For a complete working example, follow Your first AERONET query.