anise 0.9.4

ANISE provides a toolkit and files for Attitude, Navigation, Instrument, Spacecraft, and Ephemeris data. It's a modern replacement of NAIF SPICE file.

ANISE (Attitude, Navigation, Instrument, Spacecraft, Ephemeris)

ANISE is a modern rewrite of the core functionalities of the NAIF SPICE toolkit with enhanced features and ease of use, leveraging Rust's safety and speed.

Introduction

In the realm of space exploration, navigation, and astrophysics, precise and efficient computation of spacecraft position, orientation, and time is critical. ANISE, standing for "Attitude, Navigation, Instrument, Spacecraft, Ephemeris," offers a Rust-native approach to these challenges. This toolkit provides a suite of functionalities including but not limited to:

• Universal Loading: Seamlessly load SPK, BPC, PCK, FK, TPC, and the new LKA (Location Kernel Anise) files.
• Rigid Body Physics: High-precision translations, rotations, and their combination (rigid body transformations).
• Parallel Querying: Query SPICE files in parallel at incredible speeds (~ 125,000 queries per second) using the many keyword.
• Declarative Analysis: Build complex mathematical vector and scalar expressions (e.g., angles, projections, orbital elements) and evaluate them efficiently over time series.
• Event Finding: High-precision search for eclipses, AOS/LOS, and orbital events (periapsis/apoapsis).
• Hifitime Integration: Comprehensive time system conversions using the hifitime library (TT, TAI, ET, TDB, UTC, GPS, etc.).
• CCSDS OEM: ANISE supports reading, writing, and interpolating CCSDS OEM files, including their covariance, in the inertial, RIC, RCN, and VNC frames.

ANISE stands validated against the traditional SPICE toolkit, ensuring accuracy and reliability, with translations achieving machine precision (2e-16) and rotations presenting minimal error.

Why ANISE?

A quick comparison with the traditional CSPICE toolkit:

Feature	CSPICE (Toolkit)	ANISE
Thread Safety	No (Global state/locks)	Guaranteed (Rust ownership)
Performance	Single-threaded	Parallel (many queries)
Math Validation	Runtime errors	Type-safe Frame checks
API Style	Procedural (Integer IDs)	Object-Oriented / Pythonic
Serialization	None	S-Expressions (Cloud Native)

ANISE supports many SPICE kernels. Binary kernels are supported as-is, while text kernels must be transformed into their ANISE equivalent (usually .pca or .lka files).

Kernel Type	Supported	Note
BSP / SPK	✅	Supported as-is
BPC	✅	Supported as-is
TK (Text)	🔄	Supported after transformation
LSK	🔄	Supported after transformation
GM	🔄	Supported after transformation
CK	❌	Yet to be supported
SCLK	❌	Yet to be supported
DSK	❌	Yet to be supported
IK / EK	❌	Yet to be supported

Supported SPK Types

For more details on SPK types, refer to the NAIF SPK Required Reading. The following table summarizes the types supported by ANISE.

SPK Type	Interpolation	NAIF SPICE	ANISE	Typically used in
Type 1	Modified Differences	✅	✅	NASA internal trajectory tools (e.g. DPTRAJ)
Type 2	Chebyshev Triplet	✅	✅	Planetary ephemerides (e.g., JPL DE series)
Type 3	Chebyshev Sextuplet	✅	✅	Planetary ephemerides (e.g., JPL DE series)
Type 5	Discrete states	✅	❌	Two-body propagation
Type 8	Lagrange (Equal Step)	✅	🧪	Uncommon spacecraft trajectories from numerical integration
Type 9	Lagrange (Unequal Step)	✅	✅	Spacecraft trajectories from numerical integration
Type 10	Space Command TLE	✅	❌	Please don't use TLEs, a punch-card format (no joke)
Type 12	Hermite (Equal Step)	✅	🧪	Uncommon spacecraft trajectories from numerical integration
Type 13	Hermite (Unequal Step)	✅	✅	Spacecraft trajectories from numerical integration
Type 14	Chebyshev Sextuplet (Unequal Step)	✅	❌	Never seen in the wild, probably good for slow-moving objects
Type 15	Precessing Conic Propagation	✅	❌	Precessing conic elements
Type 17	Equinoctial Elements	✅	❌	NAIF docs
Type 18	ESOC/DDID Hermite/Lagrange Interpolation	✅	❌	NAIF docs
Type 19	ESOC/DDID Piecewise Interpolation	✅	❌	NAIF docs
Type 20	Chebyshev Triplet (velocity only)	✅	❌	NAIF docs
Type 21	Extended Modified Difference Arrays	✅	❌	NAIF docs

Note: 🧪 means the SPK type is supported but no public SPK of that type could be found to validate the implementation. Please provide one if you have one!

Features

• High Precision: Matches SPICE to machine precision in translations and minimal errors in rotations.
• Analysis Engine: A new declarative system to define engineering reports and events using S-Expressions, separating the definition of a calculation from its execution.
  • Cloud Native: S-Expressions are serializable, allowing you to define complex queries on a client and execute them safely on remote workers without arbitrary code execution risks.
• Ground Station Management: First-class support for Location objects, terrain masks, and visibility calculations.
• Rust Efficiency: Harnesses the speed and safety of Rust for space computations.
• Multi-threaded: ANISE is designed for modern hardware. Forget about mutexes and race conditions; ANISE guarantees thread safety.
• Frame Safety: ANISE checks that all frame translations or rotations are physically valid before performing any computation.

Architecture

graph TD
    Kernels[SPK / BPC / PCK / LKA Files] -->|Load| Almanac
    Almanac -->|Provide| Frames[Frame System]
    Almanac -->|Query| Ephemeris
    User -->|Define| Spec[Analysis Spec / Events]
    Spec -->|Evaluate| Almanac
    Almanac -->|Result| Report[Pandas/Polars DataFrame / Events]

Tutorials

• 01 - Querying SPK files
• 02 - Loading remote and local files (MetaAlmanac)
• 03 - Defining and working with the orbit structure
• 04 - Computing azimuth, elevation, and range data (AER)
• 05 - Using frame kernels and text planetary kernels
• 06 - Sun probe Earth angle
• 07 - Analysis
• 08 - OEM and Covariance

Note: The tutorials can be viewed in read-only form on the Github repo.

Usage

ANISE is available on PyPI with pre-built wheels for Linux, macOS (Intel/Silicon), and Windows.

Start by adding anise to your project: pip install anise.

Note: for the very latest usage examples, refer to the Python tests.

1. Basic Navigation & Transformations

The core of ANISE is the Almanac, which manages frames and ephemerides.

from anise import Almanac
from anise.astro import Orbit
from anise.constants import Frames
from anise.time import Epoch
from pathlib import Path

# Load your kernels (BSP, PCK, etc.)
# Note: Almanac functions are immutable; they return a NEW Almanac instance.
data_path = Path("../data")
almanac = Almanac(str(data_path.joinpath("de440s.bsp")))
almanac = almanac.load(str(data_path.joinpath("pck08.pca")))

# Access Frame properties directly
eme2k = almanac.frame_info(Frames.EME2000)
print(f"Earth GM: {eme2k.mu_km3_s2()} km³/s²")

# Define an Orbit state
epoch = Epoch("2021-10-29 12:34:56 TDB")
orig_state = Orbit.from_keplerian(
    8_191.93,   # SMA (km)
    1e-6,       # Eccentricity
    12.85,      # Inclination (deg)
    306.614,    # RAAN (deg)
    314.19,     # Arg Peri (deg)
    99.887_7,   # True Anomaly (deg)
    epoch,
    eme2k,
)

# Transform to a new frame (e.g., Earth Fixed / ITRF)
# Note: If high-precision Earth orientation files (BPC) aren't loaded, 
# this uses the IAU approximation.
state_itrf93 = almanac.transform_to(orig_state, Frames.IAU_EARTH_FRAME)

print(f"Lat: {state_itrf93.latitude_deg():.4f} deg")
print(f"Lon: {state_itrf93.longitude_deg():.4f} deg")
print(f"Alt: {state_itrf93.height_km():.4f} km")

2. Analysis & Event Finding

ANISE allows you to build declarative expressions for scalars and vectors. These are optimized in Rust and can be used to generate reports or find events (like eclipses) efficiently. These calculations are 500x faster than Ansys STK.

import anise.analysis as analysis
from anise import Almanac
from anise.constants import Frames, Orientations
from anise.astro improt Frame
from anise.time import Epoch, Unit

# Define the state we want to analyze (e.g., LRO orbiting the Moon)
lro_state = analysis.StateSpec(
    target_frame=analysis.FrameSpec.Loaded(Frame(-85, Orientations.J2000)),  # -85 is the LRO ID
    observer_frame=analysis.FrameSpec.Loaded(Frames.MOON_J2000),
    ab_corr=None,
)

# 1. Define an Event: Find when the Sun sets (elevation < 0) as seen by LRO
sun_set_event = analysis.Event(
    analysis.ScalarExpr.SunAngle(observer_id=-85), # ID -85
    analysis.Condition.LessThan(90.0), # Angle > 90 deg means sun is behind horizon
    Unit.Second * 0.5, # Precision (Units handled automatically)
    ab_corr=None,
)

# 2. Define a Report: Calculate altitude and beta angle over time
report_spec = analysis.ReportScalars([
    (analysis.ScalarExpr.Norm(analysis.VectorExpr.Radius(lro_state)), "Dist (km)"),
    (analysis.ScalarExpr.BetaAngle(), "Beta (deg)"),
], lro_state)

# Load data and execute
almanac = Almanac("../data/de440s.bsp").load("../data/lro.bsp")
start = Epoch("2025-01-01 12:00:00 UTC")
end = start + Unit.Day * 1

# Find all sunset intervals
sunset_arcs = almanac.report_event_arcs(lro_state, sun_set_event, start, end)
print(f"Found {len(sunset_arcs)} sunset intervals.")

# Generate data for the report
from anise.time import TimeSeries
series = TimeSeries(start, end, Unit.Minute * 10, inclusive=True)
data = almanac.report_scalars(report_spec, series)
# 'data' is a dictionary keyed by epoch strings, ready for pandas/plotting

3. Ground Stations

You can create and save Location Kernels (.lka) containing ground station coordinates and terrain masks.

from anise.astro import Location, TerrainMask, FrameUid

# Define a station with a visibility mask
mask = [TerrainMask(0.0, 5.0), TerrainMask(35.0, 10.0)] # Azimuth, Min Elevation
dss65 = Location(
    40.427, 4.250, 0.834, # Lat, Lon, Height (km)
    FrameUid(399, 399),   # On Earth
    mask,
    terrain_mask_ignored=False
)

# Calculate visibility
# (See tutorials for full implementation of visibility arcs)

Development

1. Install maturin, e.g. via pipx as pipx install maturin
2. Create a virtual environment: cd anise/anise-py && python3 -m venv .venv
3. Jump into the virtual environment and install patchelf for faster builds: pip install patchelf, and pytest for the test suite: pip install pytest
4. Run maturin develop to build the development package and install it in the virtual environment
5. Finally, run the tests python -m pytest

To run the development version of ANISE in a Jupyter Notebook, install ipykernels in your virtual environment.

1. pip install ipykernel
2. Now, build the local kernel: python -m ipykernel install --user --name=.venv
3. Then, start jupyter notebook: jupyter notebook
4. Open the notebook, click on the top right and make sure to choose the environment you created just a few steps above.

Generating the pyi type hints

Type hints are extremely useful for Python users. Building them is a bit of manual work.

1. maturin develop to build the latest library
2. python generate_stubs.py anise anise.pyi builds the top level type hints
3. Repeat for all submodules: utils, time, astro, astro.constants, rotation, analysis writing to a new file each time:
   1. python generate_stubs.py anise.astro anise.astro.pyi
   2. python generate_stubs.py anise.time anise.time.pyi
   3. python generate_stubs.py anise.astro.constants anise.astro.constants.pyi
   4. python generate_stubs.py anise.utils anise.utils.pyi
   5. python generate_stubs.py anise.rotation anise.rotation.pyi
   6. python generate_stubs.py anise.analysis anise.analysis.pyi
4. Final, concat all of these new files back to anise.pyi since that's the only one used by maturin.

Citation

If you use ANISE in your research, please cite it as follows:

@software{nyx_space_anise_2025,
  author = {Christopher Rabotin},
  title = {ANISE: Attitude, Navigation, Instrument, Spacecraft, Ephemeris},
  year = {2025},
  publisher = {Nyx Space},
  url = {https://github.com/nyx-space/anise}
}