beacon-api 1.3.2

Python interface wrapper for interacting with Beacon Nodes.

Beacon Data Lake Python API

Python wrapper for interacting with the Beacon Data Lake API. It discovers tables and datasets, inspects schemas, and ships a composable query builder that streams results straight into Pandas, GeoPandas, xarray, and on-disk formats such as (Geo)Parquet, NetCDF, and Zarr.

The full documentation lives at https://maris-development.github.io/beacon-py/.

Installation

pip install beacon-api

Beacon API supports Python 3.10+.

Quick start — connect

The examples below run against the public World Ocean Database (WOD) node at https://beacon-wod.maris.nl, so you can paste them straight into a notebook. Always pass a user_agent that identifies your application (and ideally a contact) so requests can be attributed on shared/public nodes.

from beacon_api import Client

client = Client(
    "https://beacon-wod.maris.nl",
    user_agent="my-app/1.0 ([e-mail])",
    # jwt_token="<bearer token>",        # for protected nodes
    # basic_auth=("user", "pass"),        # or HTTP basic auth
)

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

Getting started: SQL

Already have SQL? Run it directly and materialize the result as a DataFrame:

df = client.sql_query(
    """
    SELECT lon, lat, z, time, Temperature, Salinity
    FROM "default"
    WHERE time BETWEEN '2020-01-01T00:00:00' AND '2020-02-01T00:00:00'
    """
).to_pandas_dataframe()

print(df.head())

Getting started: JSON query builder

Prefer a fluent, typed builder? Start from a table and chain selects and filters. The builder and SQL paths share the same output helpers (to_pandas_dataframe(), to_parquet(), …).

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

df = (
    wod
    .query()
    .add_select_column("lon", alias="longitude")
    .add_select_column("lat", alias="latitude")
    .add_select_column("z", alias="depth")
    .add_select_column("time")
    .add_select_column("Temperature")
    .add_select_column("Salinity")
    .add_range_filter("time", "2020-01-01T00:00:00", "2020-02-01T00:00:00")
    .to_pandas_dataframe()
)

print(df.head())

Going further

Explore tables & schemas

list_tables() returns DataTable helpers that already know their description, type, and schema:

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

print(wod.get_table_description())
schema = wod.get_table_schema_arrow()  # pyarrow.Schema
for field in schema:
    print(field.name, field.type)

# get_table_schema() instead returns a plain dict[str, type]
print(wod.get_table_schema())

See Working with tables.

Datasets — query files directly

On Beacon ≥ 1.4.0, list_datasets() surfaces file-backed resources you can query without going through a logical table:

datasets = client.list_datasets(pattern="**/*.parquet", limit=10)
first = next(iter(datasets.values()))

print(first.get_file_name(), first.get_file_format())
df = first.query().add_select_column("lon").add_select_column("lat").to_pandas_dataframe()

See Working with datasets.

More complex queries

The JSON builder supports range/equality/null/geospatial filters, boolean combinations, distinct, sorting, and a range of output formats:

df = (
    wod
    .query()
    .add_select_column("lon")
    .add_select_column("lat")
    .add_select_column("time")
    .add_select_column("Temperature")
    .add_range_filter("time", "2020-01-01T00:00:00", "2020-06-30T23:59:59")
    .add_range_filter("z", 0, 50)
    .add_is_not_null_filter("Temperature")
    .add_bbox_filter("lon", "lat", bbox=(-20, 40, -10, 55))
    .add_sort("time", ascending=True)
    .to_pandas_dataframe()
)

See Querying the Beacon Data Lake for the full builder reference and export helpers (to_geo_pandas_dataframe, to_parquet, to_netcdf, to_zarr, …).

Streaming large results

For result sets too large to buffer, sql_query_streaming() returns a PyArrow RecordBatchStreamReader (requires Beacon ≥ 1.5.0) that you can consume batch by batch:

reader = client.sql_query_streaming('SELECT lon, lat, z, time, Temperature FROM "default"')
for batch in reader:
    # batch is a pyarrow.RecordBatch
    print(batch.num_rows)

Issues

If you encounter any issues or have feature requests, please report them on the GitHub Issues page.

Development

This project is under active development. Contributions are welcome!

To generate the typings for the API, run:

stubgen beacon_api -o .

To build the wheel package, run:

python -m build