arandu

arandu is a compact Python interface to the Arandu broker data served by ADSS. It returns query results as pandas DataFrames and keeps raw SQL available when needed.

Install

pip install -e '.[cutouts]'

Quick start

from arandu import AranduClient

client = AranduClient()
objects = client.objects(limit=100)
sources = client.sources(dia_object_id=12345)

AranduClient() prompts for credentials. Alternatively, set ARANDU_USERNAME and ARANDU_PASSWORD and use AranduClient.from_env(). The optional ARANDU_ADSS_BASE_URL defaults to https://ai-scope.cbpf.br.

from arandu import download_cutout

image = download_cutout(alert, "science")

Cutouts require the cutouts extra. alert may contain a published URL at payload.cutouts; otherwise the standard broker cutout URL is used.

Download selections

All spatial coordinates and radii are in degrees. Date selections use the Roman source table's midpoint_mjd_tai (MJD TAI).

# One DataFrame.
sources = client.download("arandu.dia_source", limit=100_000)

# For a large result, split the indexed MJD range into daily requests.
# Each interval is [start, end), so no rows are duplicated.
for batch in client.sources_by_date_batches(60000, 60010, interval_days=1):
    print(len(batch))

cone = client.cone_search(
    "arandu.dia_object", ra=277.48425, dec=13.5576, radius_deg=5 / 60
)

polygon = client.polygon_search(
    "arandu.dia_object",
    [(277.4, 13.5), (277.6, 13.5), (277.6, 13.7), (277.4, 13.7)],
)

sources = client.sources_by_date(60000, 60001)

The spatial helpers generate AstroQL cone(...) and polygon(...) clauses. Date batches constrain midpoint_mjd_tai directly and do not use OFFSET or an expensive global ORDER BY.