msts-micropython

Geocentric lunar ephemerides from a Unix timestamp: the Moon's ecliptic position, distance, and phase.

This is a Python port of msts, the OCaml library, rewritten for MicroPython on resource-constrained microcontrollers. It implements the same algorithm with the same accuracy bounds. Use this version on-device, and the OCaml original wherever a full OCaml toolchain is available.

The implementation follows Paul Schlyter's algorithm: Keplerian orbital elements propagated from a fixed epoch with empirical perturbation corrections.

The library depends only on the standard math module. It performs no I/O, reads no files, keeps no global state, and is fully deterministic. compute is total for finite inputs; invalid arguments raise ValueError naming the offending parameter.

Results are validated against JPL Horizons (DE441) over the interval 1900-01-01 through 2100-12-31.

Guaranteed error bounds:

• 2° in ecliptic longitude
• 0.5° in ecliptic latitude
• 1% in distance
• 0.02 in illuminated fraction

Typical errors are substantially smaller.

Requirements

• MicroPython, or
• CPython 3

Only the standard math module is required.

Installation

MicroPython

Either:

• Install micropython-msts from Tools → Manage packages… in Thonny, or
• Install directly from this repo, which doesn't depend on PyPI's package-search index:

mpremote mip install https://codeberg.org/duras/msts-micropython/raw/branch/main/

• Or copy msts.py onto the target device by hand, for example:

mpremote cp msts.py :

Or simply place it alongside your own modules.

CPython

Install from PyPI:

pip install micropython-msts

Example

import msts

# J2000.0 = 2000-01-01T12:00:00Z
e = msts.compute(946728000.0)

print("%s  %.1f%%  %.0f km" % (
    msts.phase_name_to_string(e.phase.name),
    e.phase.illuminated_fraction * 100.0,
    e.position.distance_km,
))

Example uses

compute takes a Unix timestamp and returns a small result structure. It performs no I/O and allocates only three small objects, making it inexpensive enough to call on every wake cycle of a battery-powered device.

Some applications:

• Clock or watch firmware — moon-phase complications for MicroPython-based clocks and smartwatches.
• E-ink or OLED displays — ESP32 or RP2040 boards that wake on a timer, compute the current phase, and update a display without network access.
• Garden automation — irrigation or planting reminders keyed to the lunar phase alongside existing sensor data.
• Outdoor lighting — dim or disable solar- or LoRa-powered lighting near the full moon when ambient light is already higher.
• Education — classroom demonstrations of basic orbital mechanics by varying the timestamp and observing the resulting position and phase.
• Astrophotography helpers — approximate moon phase and brightness for scheduling. Not suitable where precise rise/set times or arcsecond-level accuracy are required.

Differences from the OCaml original

• phase_name is represented by small integer constants (msts.NEW_MOON, msts.WAXING_CRESCENT, ...) instead of an OCaml variant type or Python enum, since MicroPython has no standard enum module.
• Invalid_argument becomes ValueError.
• Field names, function names, formulas, constants, and documented accuracy bounds otherwise match the OCaml implementation one-for-one.

License

ISC. See LICENSE.