Skip to content

Beacon API Python SDK

The Beacon Python SDK is a client for Beacon Data Lakes. It discovers datasets, inspects schemas, and ships a composable query builder that can stream results into familiar scientific Python tools.

Why use the SDK?

  • One client for every Beacon Node – the Client wires authentication headers, probes /api/health, and exposes helpers such as list_tables() and list_datasets().
  • Discoverability built-in – DataTable and Dataset helpers return Arrow schemas, table descriptions, and file metadata so you always know which columns exist before writing a query.
  • Chainable JSON/SQL query builders – start from a table or dataset, add selects, filters, distinct clauses, or geospatial predicates, and export to Pandas, GeoPandas, Dask, xarray, or on-disk formats such as (Geo)Parquet, Arrow IPC, NetCDF, NdNetCDF, CSV, Zarr, or ODV.
  • Typed, well-documented API surface – the docs you are reading mirror the public classes (Client, DataTable, Dataset, JSONQuery, SQLQuery, …) so editors and notebooks surface the same guidance.

Beacon Data Lake platform

Looking for the platform documentation itself? Head over to the Beacon Data Lake docs.

Quick start

from beacon_api import Client

client = Client(
    "https://beacon.example.com",
    jwt_token="<optional bearer token>",
)

client.check_status()  # probes /api/health and prints the Beacon version

tables = client.list_tables()
stations = tables["default"]

df = (
    stations
    .query()
    .add_select_columns([
        ("LONGITUDE", None),
        ("LATITUDE", None),
        ("JULD", None),
        ("TEMP", "temperature_c"),
    ])
    .add_range_filter("JULD", "2024-01-01T00:00:00", "2024-12-31T23:59:59")
    .to_pandas_dataframe()
)

Core concepts

Client

Manages the HTTP session, authentication headers, and compatibility checks. Use get_server_info() to inspect the Beacon version and extensions, list_tables() to enumerate logical collections, or list_datasets() (Beacon ≥ 1.4.0) to work with direct file paths.

Tables and datasets

Tables (instances of DataTable) represent logical collections backed by one or more datasets. They expose helpers such as get_table_schema() or subset() to quickly explore spatial/temporal windows. Datasets expose similar helpers but start from a known file/URI and can construct a query via Dataset.query().

Query builders

JSONQuery powers the fluent builder: select columns, coalesce values, add range/equality/geospatial filters, sorting, distinct clauses, or call explain() to see the server-side plan. Prefer Client.sql_query() when you already have SQL.

Rich outputs

Every query inherits the BaseQuery output helpers. Stream into to_pandas_dataframe(), to_geo_pandas_dataframe(), to_dask_dataframe(), to_xarray_dataset(), or write datasets using to_parquet(), to_nd_netcdf(), to_zarr(), to_odv() and more.

Where to next?

  • Installation – supported Python versions and optional extras.
  • Getting started – end-to-end walkthrough from connecting a client to exporting a DataFrame.
  • Using Beacon – deep dives into tables, datasets, and the JSON query builder.
  • Working with datasets – jump straight from file paths/URIs to a query builder.
  • API reference – auto-generated API docs straight from the SDK source.