====== QSOnic

Lightining-fast continuum fitting

.. image:: https://joss.theoj.org/papers/10.21105/joss.06373/status.svg :target: https://doi.org/10.21105/joss.06373

.. image:: https://img.shields.io/pypi/v/qsonic?color=blue :target: https://pypi.org/project/qsonic

.. image:: https://img.shields.io/badge/Source-qsonic-red :target: https://github.com/p-slash/qsonic

.. image:: https://github.com/p-slash/qsonic/actions/workflows/testing.yml/badge.svg :target: https://github.com/p-slash/qsonic/actions/workflows/testing.yml :alt: Tests Status

.. image:: https://readthedocs.org/projects/qsonic/badge/?version=latest :target: https://qsonic.readthedocs.io/en/latest/?badge=latest :alt: Documentation Status

QSOnic is an MPI-parallelized, highly optimized quasar continuum fitting package for DESI built on the same algorithm as picca <https://github.com/igmhub/picca>_, but faster. It also provides an efficient API to read DESI quasar spectra. If you use this software, please cite the article in the Journal of Open Source Software (JOSS).

The key differences

• You can use any desired input continuum from another continuum prediction method as input continuum model.
• Coadding of spectrograph arms can be performed after continuum fitting or disabled entirely.
• Continuum is multiplied by a fiducial mean flux when provided.
• You can pass fiducial var_lss (column VAR_LSS) and mean flux (column MEANFLUX) for observed wavelength LAMBDA in STATS extention of a FITS file. Wavelength should be linearly and equally spaced. This is the same format as rawio output from picca, except VAR column in picca is the variance on flux not deltas. We break away from that convention by explicitly requiring variance on deltas in a new column.
• If no fiducial is passed, we fit only for var_lss (no eta fitting by default). Eta fitting can be enabled by passing --var-fit-eta. A covariance matrix is calculated based on delete-one Jackknife over subsamples between var_pipe and var_obs data points, and used in var_lss, eta fitting if --var-use-cov passed (soft recommendation to always enable this).
• Internal weights for continuum fitting and coadding are based on smoothed IVAR, and output WEIGHT is based on this smoothed ivar. This smoothing can be turned off.
• Chi2 information as well as best fits are saved in continuum_chi2_catalog.fits. Chi2 is calculated using smooth ivar and var_lss, and does not subtract sum of ln(weights).

Similarities

• Delta files are the same. CONT column is mean flux times continuum even when fiducial mean flux is passed.
• MEANSNR in header file and chi2 catalog is average of flux times square root of positive ivar values. Header values are per arm, but catalog values are the average over all arms.
• Eta fitting does not rescale IVAR output. Pipeline noise will be modified with explicit calibration option.