adam_core: ADAM Core Utilities

A Python package by the Asteroid Institute, a program of the B612 Foundation

adam_core is used by a variety of library and services at the Asteroid Institute. Sharing these common classes, types, and conversions amongst our tools ensures consistency and accuracy.

Installation

ADAM Core is available on PyPI

pip install adam_core

Astropy, Astroquery, Healpy, and plotting libraries are explicit optional providers rather than default runtime dependencies. Install only the bridge you need:

pip install "adam_core[astropy]"      # Astropy Time and UT1/IERS bridges
pip install "adam_core[legacy-sbdb]"  # Astroquery compatibility workflows
pip install "adam_core[healpix]"      # Healpy-backed public helpers
pip install "adam_core[jax]"          # Historical explicit JAX bridge
pip install "adam_core[plots]"        # Plotting and date-axis formatting

Native wheels support CPython 3.11-3.13 on manylinux 2.17+ x86-64/AArch64 and macOS Apple silicon/Intel. Windows is deferred because the required libassist-sys 1.2.1 acceptance stack depends on upstream ASSIST's POSIX memory mapping; musllinux is also unsupported. Python installs reuse kernel data from the active environment; pure-Rust consumers use override -> installed-Python -> cache -> checksummed wheel resolution, with ADAM_CORE_KERNEL_OFFLINE=1 disabling downloads.

ASSIST propagation is provided by the separate adam-assist distribution as adam_assist.ASSISTPropagator. adam-assist owns the orchestration layer and consumes libassist-sys and librebound-sys directly.

Usage

Orbits

To define an orbit:

from adam_core.coordinates import KeplerianCoordinates
from adam_core.coordinates import Origin
from adam_core.orbits import Orbits
from adam_core.time import Timestamp

keplerian_elements = KeplerianCoordinates.from_kwargs(
    time=Timestamp.from_mjd([59000.0], scale="tdb"),
    a=[1.0],
    e=[0.002],
    i=[10.],
    raan=[50.0],
    ap=[20.0],
    M=[30.0],
    origin=Origin.from_kwargs(code=["SUN"]),
    frame="ecliptic"
)
orbits = Orbits.from_kwargs(
    orbit_id=["1"],
    object_id=["Test Orbit"],
    coordinates=keplerian_elements.to_cartesian(),
)

Note that internally, all orbits are stored in Cartesian coordinates. Cartesian coordinates do not have any singularities and are thus more robust for numerical integration. Any orbital element conversions to Cartesian can be done on demand by calling to_cartesian() on the coordinates object.

The underlying orbits class is 2 dimensional and can store elements and covariances for multiple orbits.

from adam_core.coordinates import KeplerianCoordinates
from adam_core.coordinates import Origin
from adam_core.orbits import Orbits
from adam_core.time import Timestamp

keplerian_elements = KeplerianCoordinates.from_kwargs(
    time=Timestamp.from_mjd([59000.0, 60000.0], scale="tdb"),
    a=[1.0, 3.0],
    e=[0.002, 0.0],
    i=[10., 30.],
    raan=[50.0, 32.0],
    ap=[20.0, 94.0],
    M=[30.0, 159.0],
    origin=Origin.from_kwargs(code=["SUN", "SUN"]),
    frame="ecliptic"
)
orbits = Orbits.from_kwargs(
    orbit_id=["1", "2"],
    object_id=["Test Orbit 1", "Test Orbit 2"],
    coordinates=keplerian_elements.to_cartesian(),
)

Orbits can be easily converted to a pandas DataFrame:

orbits.to_dataframe()  
  orbit_id     object_id  coordinates.x  coordinates.y  coordinates.z  coordinates.vx  coordinates.vy  coordinates.vz  coordinates.time.days  coordinates.time.nanos                      coordinates.covariance.values coordinates.origin.code  
0        1  Test Orbit 1      -0.166403       0.975273       0.133015       -0.016838       -0.003117        0.001921                  59000                       0  [nan, nan, nan, nan, nan, nan, nan, nan, nan, ...                     SUN  
1        2  Test Orbit 2       0.572777      -2.571820      -1.434457        0.009387        0.002900       -0.001452                  60000                       0  [nan, nan, nan, nan, nan, nan, nan, nan, nan, ...                     SUN

Orbits can also be defined with uncertainties.

import numpy as np
from adam_core.coordinates import KeplerianCoordinates
from adam_core.coordinates import Origin
from adam_core.coordinates import CoordinateCovariances
from adam_core.orbits import Orbits
from adam_core.time import Timestamp

keplerian_elements = KeplerianCoordinates.from_kwargs(
    time=Timestamp.from_mjd([59000.0], scale="tdb"),
    a=[1.0],
    e=[0.002],
    i=[10.],
    raan=[50.0],
    ap=[20.0],
    M=[30.0],
    covariance=CoordinateCovariances.from_sigmas(
        np.array([[0.002, 0.001, 0.01, 0.01, 0.1, 0.1]])
    ),
    origin=Origin.from_kwargs(code=["SUN"]),
    frame="ecliptic"
)

orbits = Orbits.from_kwargs(
    orbit_id=["1"],
    object_id=["Test Orbit with Uncertainties"],
    coordinates=keplerian_elements.to_cartesian(),
)
orbits.to_dataframe()  
  orbit_id                      object_id  coordinates.x  coordinates.y  coordinates.z  coordinates.vx  coordinates.vy  coordinates.vz  coordinates.time.days  coordinates.time.nanos                      coordinates.covariance.values coordinates.origin.code  
0        1  Test Orbit with Uncertainties      -0.166403       0.975273       0.133015       -0.016838       -0.003117        0.001921                  59000                       0  [6.654136535278775e-06, 1.2935845684776213e-06...                     SUN

The covariance matrices can be extracted in matrix form by using the .to_matrix() method:

orbits.coordinates.covariance.to_matrix()

Similarly, if you just want to access the orbital elements you can do the following:

orbits.coordinates.values

To query orbits from JPL Horizons:

from adam_core.orbits.query import query_horizons
from adam_core.time import Timestamp

times = Timestamp.from_mjd([60000.0], scale="tdb")
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_horizons(object_ids, times)

To query orbits from JPL SBDB:

from adam_core.orbits.query import query_sbdb

object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_sbdb(object_ids)

Orbital Element Conversions

Orbital elements can be accessed via the corresponding attribute. All conversions, including covariances, are done on demand and stored.

# Cartesian Elements
orbits.coordinates

# To convert to other representations
cometary_elements = orbits.coordinates.to_cometary()
keplerian_elements = orbits.coordinates.to_keplerian()
spherical_elements = orbits.coordinates.to_spherical()

Propagator

The propagator class in adam_core provides a generalized interface to the supported orbit integrators and ephemeris generators. The propagator class is designed to be used with the Orbits class and can handle multiple orbits and times.

You will need to install either adam-assist, or another compatible propagator in order to use propagation, ephemeris generation, or impact analysis.

Propagation

To propagate orbits with ASSIST (here we grab some orbits from Horizons first):

import numpy as np
from astropy import units as u

from adam_core.orbits.query import query_horizons
from adam_assist import ASSISTPropagator
from adam_core.time import Timestamp

# Get orbits to propagate
initial_time = Timestamp.from_mjd([60000.0], scale="tdb")
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_horizons(object_ids, initial_time)

# initialize the propagator
propagator = ASSISTPropagator()

# Define propagation times
times = initial_time.from_mjd(initial_time.mjd() + np.arange(0, 100))

# Propagate orbits! This function supports multiprocessing for large
# propagation jobs.
propagated_orbits = propagator.propagate_orbits(
    orbits,
    times,
    chunk_size=100,
    max_processes=1,
)

Ephemeris Generation

Ephemeris generation requires a compatible propagator. adam-assist supplies the supported high-precision N-body implementation and maps propagated covariance matrices to the sky plane.

import numpy as np
from astropy import units as u

from adam_assist import ASSISTPropagator
from adam_core.observers import Observers
from adam_core.orbits.query import query_horizons
from adam_core.time import Timestamp

# Get orbits to propagate
initial_time = Timestamp.from_mjd([60000.0], scale="tdb")
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_horizons(object_ids, initial_time)

propagator = ASSISTPropagator()

# Define a set of observers and observation times
times = Timestamp.from_mjd(initial_time.mjd() + np.arange(0, 100))
observers = Observers.from_code("I11", times)

# Generate ephemerides! This function supports multiprocessing for large
# propagation jobs.
ephemeris = propagator.generate_ephemeris(
    orbits,
    observers,
    chunk_size=100,
    max_processes=1
)

Low-level APIs

State Vectors from Development Ephemeris files

Getting the heliocentric ecliptic state vector of a DE440 body at a given set of times (in this case the barycenter of the Jovian system):

import numpy as np

from adam_core.coordinates import OriginCodes
from adam_core.utils import get_perturber_state
from adam_core.time import Timestamp

states = get_perturber_state(
    OriginCodes.JUPITER_BARYCENTER,
    Timetamp.from_mjd(np.arange(59000, 60000), scale="tdb"),
    frame="ecliptic",
    origin=OriginCodes.SUN,
)

2-body Propagation

adam_core also has 2-body propagation functionality. To propagate any orbit with 2-body dynamics:

import numpy as np
from astropy import units as u

from adam_core.orbits.query import query_sbdb
from adam_core.dynamics import propagate_2body
from adam_core.time import Timestamp

# Get orbit to propagate
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_sbdb(object_ids)

# Define propagation times
times = Timestamp.from_mjd(np.arange(59000, 60000), scale="tdb")

# Propagate orbits with 2-body dynamics
propagated_orbits = propagate_2body(
    orbits,
    times
)

2-body Ephemeris Generation

This package also has functionality to generate ephemerides for a set of orbits. We do not recommend you use this with 2-body propagated orbits as it will not be accurate for more than a few days. However, if you used a N-body propagator such as ASSIST, you can feed in the propagated orbits to this function to generate ephemerides. We call the ephemeris generator 2-body because the light-time correction is applied using a 2-body propagator.

The ephemeris generator can map covariances directly to the sky plane. To do this, propagate the covariance matrices with the orbits. This is done by passing covariance=True to the propagator. The ephemeris generator will then automatically map the propagated covariance matrices to the sky-plane.

import numpy as np
from astropy import units as u

from adam_assist import ASSISTPropagator
from adam_core.observers import Observers
from adam_core.orbits.query import query_sbdb
from adam_core.dynamics import generate_ephemeris_2body
from adam_core.time import Timestamp

# Get orbits to propagate
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_sbdb(object_ids)

propagator = ASSISTPropagator()

# Define a set of observers and observation times
times = Timestamp.from_mjd(np.arange(59000, 60000), scale="tdb")
observers = Observers.from_code("I11", times)

# Propagate orbits with ASSIST (including covariances)
propagated_orbits = propagator.propagate_orbits(
    orbits,
    times,
    chunk_size=100,
    max_processes=1,
    covariance=True,
)

# Now generate ephemerides with the 2-body ephemeris generator
ephemeris = generate_ephemeris_2body(
    propagated_orbits,
    observers,
)

Gravitational parameter

Both the 2-body propagation and 2-body ephemeris generation code will determine the correct graviational parameter to use from each orbit's origin.

To see the gravitational parameter used for each orbit:

from adam_core.orbits.query import query_sbdb

# Get orbit to propagate
object_ids = ["Duende", "Eros", "Ceres"]
orbits = query_sbdb(object_ids)

# Get the gravitational parameter (these will all be the same -- heliocentric)
mu = orbits.coordinates.origin.mu()

Package Structure

adam_core
├── constants.py  # Shared constants
├── coordinates   # Coordinate classes and transformations
├── dynamics      # Numerical solutions
├── orbits        # Orbits class and query utilities
└── utils         # Utility classes like Indexable or conversions like times_from_df