Fixes issue raised in desihub/desispec#1794 by printing a usage statement and returning a non-zero value when --help or -h is in the list of options provide as an argument to run_specex. This was just a matter of cleaning up and using the existing PrintHelp C++ function.
It would have been cleaner to move the argparsing entirely to the python side, and then have run_specex take multiple input options like in_psf, out_psf, first_bundle, etc. rather than having run_specex do argument parsing, but this would take a lot more time and testing than we have at the moment, so I've left it for a future update.
usage: desi_psf_fit [options]
--help,-h display usage information
--arc,-a preprocessed fits image file name (mandatory)
--in-psf input psf file name (fits or xml, can contain
only the traces, mandatory)
--out-psf output psf fits file name (mandatory)
--lamp-lines input lamp lines file name (mandatory)
--only-trace-fit only fit the trace coordinates (default is
trace,sigma,psf)
--no-trace-fit do not fit the trace coordinates (default is
trace,sigma,psf)
--no-sigma-fit do not fit the gaussian sigma (default is
trace,sigma,psf)
--no-psf-fit do not fit the psf (default is trace,sigma,psf)
--flux-hdu flux hdu in input arc fits, for unusual data format
--ivar-hdu ivar hdu in input arc fits, for unusual data format
--mask-hdu mask hdu in input arc fits, for unusual data format
--header-hdu header hdu in input arc fits, for unusual data format
--xcoord-hdu hdu of x trace legendre polynomial of wavelength
(default is extension XTRACE)
--ycoord-hdu hdu of y trace legendre polynomial of wavelength
(default is extension YTRACE)
--first-bundle first fiber bundle to fit
--last-bundle last fiber bundle to fit
--single-bundle fit a single bundle with all fibers
--first-fiber first fiber (must be in bundle)
--last-fiber last fiber (must be in bundle)
--half-size-x half size of PSF stamp (full size is 2*half_size+1)
--half-size-y half size of PSF stamp (full size is 2*half_size+1)
--psfmodel PSF model, default is GAUSSHERMITE
--positions fit positions of each spot individually after global
fit for debugging
--verbose,v turn on verbose mode (deprecated, true by default)
--quiet no info message, only warning
--debug turn on debug mode
--trace-prior-deg force equal trace coeff in bundle starting at this
degree
--lamplines lamp lines ASCII file name (def. is
$SPECEXDATA/specex_linelist_desi.txt)
--core dump core files when an exception is thrown
--gauss-hermite-deg degree of Hermite polynomials (same for x and y,
only if GAUSSHERMITE psf)
--gauss-hermite-deg2 degree of Hermite polynomials (same for x and y,
only if GAUSSHERMITE2 psf)
--legendre-deg-wave degree of Legendre polynomials along wavelength
(can be reduced if missing data)
--legendre-deg-x degree of Legendre polynomials along x_ccd (can be
reduced if missing data)
--trace-deg-wave degree of Legendre polynomials along wavelength for
fit of traces
--trace-deg-x degree of Legendre polynomials along x_ccd for fit of
traces
--psf-error psf fractional uncertainty (default is 0.01, for
weights in the fit)
--psf-core-wscale scale up the weight of pixels in 5x5 PSF core
--broken-fibers broken fibers (comma separated list)
--variance-model refit at the end with a model of the variance to
avoid Poisson noise bias
--out-psf-xml output psf xml file name
--out-spots output spots file name
--prior gaussian prior on a param : 'name' value error
--tmp_results write tmp results
--fit-psf-tails unable fit of psf tails
--fit-continuum unable fit of continuum
--nlines max # emission lines used (uses an algorithm to
select best ones
based on S/N and line coverage
Fixes issue raised in desihub/desispec#1794 by printing a usage statement and returning a non-zero value when
--helpor-his in the list of options provide as an argument torun_specex. This was just a matter of cleaning up and using the existingPrintHelpC++ function.It would have been cleaner to move the argparsing entirely to the python side, and then have run_specex take multiple input options like in_psf, out_psf, first_bundle, etc. rather than having run_specex do argument parsing, but this would take a lot more time and testing than we have at the moment, so I've left it for a future update.
I tested this and it works fine, e.g.,
results in the following printed to the terminal:
@sbailey