API reference

Leaf modules are listed explicitly rather than recursively: several are re-exported from their package __init__, and a recursive sweep documents those objects twice, which Sphinx reports as a duplicate description.

specsr.inference.pipeline	The user-facing inference pipeline: LR spectrum in, SR spectrum and redshift out.
specsr.paths	Filesystem locations used by specsr.
specsr.checkpoints	On-demand model weight fetching from the Hugging Face Hub.
specsr.config	Configuration loading for the training stages.
specsr.runtime	Environment-variable overrides for the training scripts.
specsr.metrics	Residual and power-spectrum statistics used by the paper figures.
specsr.linefit	Gaussian emission-line fitting, and the S/N derived from it.
specsr.data.grid	The common wavelength grid, and flux-conserving resampling onto it.
specsr.data.augment	Augmentation of paired spectra, with explicit provenance.
specsr.data.build	Build paired training products from the raw JADES release.
specsr.data.ingest	Discovery and reading of raw JADES NIRSpec x1d products.
specsr.data.stitch	Combine the medium gratings into one high-resolution reference.
specsr.data.datasets	Torch datasets over the paired spectra products.
specsr.data.splits	Group-aware train/test splitting for the augmented spectra datasets.
specsr.models.blocks	Building blocks shared across the SR1, ZHead and SR2 architectures.
specsr.models.sr1	SR1 — the super-resolution backbone (stage 1 of 3).
specsr.models.zhead	ZHead — redshift inference from a super-resolved spectrum (stage 2 of 3).
specsr.models.sr2	SR2 — physics-informed residual refiner (stage 3 of 3).
specsr.models.lines	Rest-frame spectral feature catalogue.
specsr.training.losses	Loss functions for the SR1 and SR2 stages.
specsr.training.ztransform	Redshift normalisation and bounded decoding.
specsr.training.zhead_sources	Input representations for the redshift head.