Mimir

Mimir provides shared tools for preparing astronomical time series and computing power-density spectra for asteroseismology. It separates numerical spectrum calculation from optional archive access so that projects such as PBjam, Skuld, and Urdr can use the same tested implementation.

Mimir currently provides:

• validated NumPy-based time-series containers;
• Lomb–Scargle power spectra calculated directly with nifty-ls;
• explicit Parseval normalization, power density, and amplitude;
• spectral-window and effective independent-frequency-spacing calculations;
• optional MAST access and basic reduction through Lightkurve.

Installation

Clone the repo and do:

python -m pip install -e .

for the light-weight version, or for if you want to include download handling from MAST :

python -m pip install -e ".[mast]"

The import the module:

import mimir

Until the first PyPI release, install the latest public source with:

python -m pip install "mimir-astro @ git+https://github.com/nielsenmb/Mimir.git"

Quick start

from mimir import TimeSeries, power_spectrum, spectral_window

series = TimeSeries(time, flux, flux_err, time_unit="d", flux_unit="ppm")
spectrum = power_spectrum(time_series=series, oversampling=2)
window = spectral_window(time_series=series)

print(spectrum.frequency, spectrum.power_density)
print(window.effective_frequency_spacing)

MAST access and basic Lightkurve reduction are available with the mast extra:

from mimir import power_spectrum

spectrum = power_spectrum(
    target="KIC 8006161",
    mast_kwargs={
        "search_kwargs": {"mission": "Kepler", "exptime": 60},
        "numax": 3500,
    },
)

The numerical entry points expose three distinct input forms: time with flux for arrays, time_series for a validated TimeSeries, or target for a MAST-resolvable name. Exactly one form must be selected. Target input uses the same load_lightcurve pipeline; mast_kwargs accepts loader options, with Lightkurve search constraints nested under search_kwargs.

Example notebooks

The examples directory contains tutorials for:

• validating local arrays and calculating a normalized power spectrum;
• downloading and reducing a MAST light curve through Lightkurve; and
• interpreting spectral windows and effective frequency spacing for gapped observations; and
• comparing the PBjam and Mimir spectra of the same reduced Kepler target.

The local-array and spectral-window notebooks use deterministic synthetic data and run without network access.

Numerical conventions

Power is one-sided and normalized over the physical band through Nyquist. At the default nyquist_factor=1, the sum of the power equals the input flux variance. Power density integrates to the same variance, and amplitude is a sinusoidal semi-amplitude with the same units as the input flux. See the spectrum documentation for the complete definitions.

Mimir uses nifty-ls directly rather than Astropy's standard Lomb–Scargle backend. NumPy arrays are the public interchange format; Lightkurve is required only for archive access and reduction.

Contributing

Bug reports, numerical test cases, and focused pull requests are welcome. See CONTRIBUTING.md for the development setup and conventions. All public functions and classes use NumPy-style docstrings.

Citation

If Mimir contributes to published research, please cite the software using the metadata in CITATION.cff. Release-specific archive information can be added after the first public release.

License

Mimir is distributed under the MIT License.