Module Guide¶

This page is a high-level map of the source tree. It avoids importing heavy runtime dependencies during documentation builds.

Package Entry Points¶

Module	Responsibility
`bayRing.bayRing`	Command-line workflow: read config, load NR data, build model, run inference and post-process.
`bayRing.initialise`	Defaults, config parsing, output-directory creation and help text.
`bayRing.NR_waveforms`	Catalogue readers, NR metadata extraction, error estimation and waveform cropping.
`bayRing.QNM_utils`	QNM mode parsing, quadratic-mode parsing and QNM frequency cache setup.
`bayRing.template_waveforms`	Waveform-model wrapper used by the inference model.
`bayRing.inference`	Dynamic sampler model, priors, nested sampling, minimization and linear inversion.
`bayRing.postprocess`	Point estimates, posterior reads, waveform plots, mismatch/SNR diagnostics and corner plots.
`bayRing.waveform_utils`	Waveform alignment, PSD/ACF utilities, physical scaling and helper calculations.
`bayRing.utils`	Shared utility functions for fixed parameters, paths and local data prefixes.

Main Runtime Objects¶

NR_waveforms.NR_simulation: Loads a selected catalogue simulation, builds NR_cpx_cut and NR_cpx_err_cut, stores metadata and tracks peak-time conventions.
template_waveforms.WaveformModel: Holds model configuration and evaluates the selected template on the NR time grid. It delegates Kerr, KerrBinary and TEOBPM physics to pyRing waveform classes.
inference.Dynamic_InferenceModel: Builds a sampler-compatible class from either cpnest.model.Model or raynest.model.Model. It creates the parameter names, bounds, fixed parameter dictionary, likelihood and prior.
inference.Minimization_Algorithm: Runs bounded least squares with one or more starting points and estimates local errors from the inverse weighted Fisher matrix.
inference.KerrLinearInversion_Algorithm: Solves the weighted normal equations for free Kerr complex amplitudes.

Testing Pointers¶

The test suite currently covers utility functions, waveform utilities, source-structure expectations, template-waveform behaviour and inference helpers:

pytest tests

The linear-inversion tests use fake waveform objects so the numerical method can be checked without loading full NR data.