caelus-engine 0.9.0

Clean-room astrological ephemeris engine: planetary and lunar positions, houses, aspects, and events from published sources. The Python reference for the Caelus engine.

caelus-engine

The Python reference implementation of the Caelus astrological ephemeris engine. Clean-room, MIT-licensed, written from published sources (VSOP87D, ELP/Meeus, IAU models, JPL Horizons fits). No Swiss Ephemeris code, no bundled third-party ephemeris files.

The engine reads pre-built Chebyshev and series packs shipped with the package and runs on the Python standard library alone. It is also the reference that the TypeScript engine (caelus on npm) is pinned to by a golden conformance suite.

Install

pip install caelus-engine

The import package is astroengine:

from astroengine import Engine, BODIES

eng = Engine("full")
jd = 2451545.0  # 2000-01-01 12:00 UT
for body in BODIES:
    print(f"{body:10s} {eng.longitude(body, jd):10.5f}")

What it computes

• Planetary and lunar apparent geocentric ecliptic longitudes (VSOP87D for the planets, Meeus Ch. 47 for the Moon, Meeus Ch. 37 series for Pluto), with light-time, annual aberration, FK5 frame correction, and IAU 1980 nutation.
• 12 house systems; tropical and sidereal (7 ayanamsas).
• Aspects with configurable orbs.
• Events: rise/set/transit, longitude crossings, lunar phases, stations, Gauquelin sectors, solar and lunar eclipses.
• Fixed stars and topocentric positions.
• Derived charts: returns, secondary progressions, solar arc, composite and Davison, harmonics, antiscia, declination aspects and parallels, out-of-bounds, dignities, sect.
• Electional primitives: applying/separating aspects, solar phase (cazimi/combust/under-the-beams), planetary hours, void-of-course Moon, house placement.
• A declarative when() query engine over celestial predicates.
• A turbo tier (Turbo): segmented Chebyshev longitude packs fit to the engine for bulk scans.

Accuracy

Accuracy is calibrated against Swiss Ephemeris and validated against JPL Horizons, stated per body rather than as a blanket figure. See the validation tables at https://ephemengine.com.

Optional: the fitting toolchain

The engine runs without any third-party dependency. The data-fitting tools that regenerate the coefficient packs need numpy:

pip install "caelus-engine[fit]"

License

MIT. See LICENSE. Project home: https://ephemengine.com. Source: https://github.com/heavyblotto/caelus.