Definition of pfsReference

We have pfsReference which includes the best-fit model template (or set) of flux standards. pfsReference is generated from `calculateReferenceFlux` and then is passed to `fluxCalibrate`. The datamodel definition has been not yet fixed and is not mentioned in datamodel.txt. We propose its definition here. Although this definition can change because we are developing flux calibration, we share our thought with you and we would like to get feedback from you. We will update datamodel.txt according to the discussion here.
The biggest change from the initial assumption in the documents is that pfsReference includes the model template set of a single visit. pfsReference is initially assumed to inherit PfsSimpleSpectrum which represents a spectrum of a single object
(see
https://github.com/Subaru-PFS/datamodel/blob/92c6865d560524971530f000162ebdc59b51308b/python/pfs/datamodel/drp.py#L31 )

 
Our proposal:
 
Reference model
 
The reference model set is a product from `calculateReferenceFlux` and is stored in the pfsReference file. pfsReference is used in the `fluxCalibrate` procedure together with pfsMerged to produce flux-calibrated spectra. pfsReference should include the wavelength and flux of the reference spectra of standard stars. Optionally, the information useful in debugging for developers are included in pfsReference. We propose that pfsReference includes multiple objects in a visit like pfsMerged.
 
The reference model set is saved to:
"pfsReference-%06d.fits" % (visit,)
 
The file has several HDUs:
 
HDU #0 PDU
HDU #1 FIBERID Fiber identifier [32-bit INT] NFIBER
HDU #2 WAVELENGTH Wavelength in units of nm [FLOAT] NROW*NFIBER
HDU #3 FLUX Flux of reference models in units of nJy [FLOAT] NROW*NFIBER
HDU #4 FITFLAG Flag if the fitting was successful or not [32-bit INT] NFIBER
HDU #5 PDF PDFs and Chi-square of fitting [FLOAT] NFIBER*6 optional
HDU #6 METADATA Table of by-products [BINARY TABLE] optional
 
PDU includes the visit number.
 
HDU #4 includes the flag of whether the fitting process was successful or not. This flag also includes the causes of failure.
 
HDUs #5 and 6 are for debugging. HDU #5 includes the probability distribution functions and chi-squares of three types: broad-band SED fitting, spectral fitting, and combined one of both. HDU #6 is a binary table of by-products from the calculateReferenceFlux process, which at least includes the stellar parameters of the reference models, the estimated radial velocity, the fitting parameters to continuum, the scaling factor, and Galactic extinction.