brunosanchez / pysmp

Scene Modeling Pipeline for Sub Percent Transient Photometry

pysmp v1.0

Knobs

Data Ingestion Knobs

Inside default.config

• -r /path/to/topdir/of/images
• --filter=g: Each filter is fit individually so you must specify your filter of interest
• -p /path/to/default.param: discussed below
• -o /path/to/outdir/for/smp
• --snfilepath=/path/to/forced/photometry/smp/input/lightcurves: SMP requires forced photometry lightcurves as an input.
• --nozpt: will always calculate the zeropoint of every image containin the SN ra and dec even if they already been calculated on a past run.
• --usefake: if fakes are present in the image given by Kessler pipeline then this will handle them accordingling and propagate data for later analysis.
• --bigstarcatalog=/path/to/big/star.cat
• --snfilelist=/path/to/file/containing/all/forced/photometry/input/files.txt

Inside default.param

• mjdplus: (mjd-peakmjd) larger than this value are fixed to zero SN flux (necessary for SMP)
• mjdminus: (mjd-peakmjd) smaller than this value are fixed to zero SN flux (necessary for SMP)
• HDR_AIRMASS_NAME: AIRMASS
• HDR_PLATESCALE_NAME: PIXSCAL1
• HDR_PSF_FWHM: PSF_FWHM
• FORCERADEC: False
• FRA: 40.514263
• FDEC: -1.630960

MCMC Knobs

Inside default.param

• --doremap: will use Galsim to remap the galaxy model to the WCS grid of each nights image (SLOW but recommended for final photometry).

Inside default.param

• GALMODEL_PIXELATION_FACTOR: 2. # NO pixelation = 1.
• SN_PLUS_GALMODEL_STEPS: 500000
• SN_SHIFT_STD: 0.0001
• SN_FLUX_PROPOSAL_MULT: 2. # Proposal STD is sqrt(diffim_flux)*SN_FLUX_PROPOSAL_DIV
• GALMODEL_PROPOSAL_STD: .6

Doing a Fit

python smpshift.py --index=$1 --nozpt

Submit to your favorite batch grid system with each job containing a different --index=$1 .

Afterwards you will run the following command to regenerate lightcurves with SMP fluxes in a new directory! (dont overwrite old lightcurves!)

python addcoltoDESlightcurve.py --lcdir=/path/to/new/lc/files --resultsdir=/path/to/smp/results

Example output lightcurve file

Contains the original forced photometry lightcurve file structure with the addition of several columns

SMP_FLUX - fit flux (at zpt of 31)
SMP_FLUXERR - fit flux error (at zpt of 31)
SMP_FLUX_ZPT - zeropoint of smp flux (always 31)
SMP_FIT_ZPT - smp fit zeropoint which is used to scale all fluxes to 31
SMP_FIT_ZPT_STD - uncertainty in smp fit zeropoint which is used to scale all fluxes to 31
SMP_CHISQ - (sim_stamp - image_stamp)^2/(err)^2
SMP_SKY - fit value for the sky
SMP_SKYERR - uncertainty in the sky
SMP_FIX - 1 means SN flux was fixed to zero in smp fit
SMP_FLAG - 1 means this epoch was flagged for some reason inside smp pipeline

(just to be clear: ALL FLUXES, SKYS, SKYERRS, ETC... ARE REPORTED AT A ZEROPOINT OF 31)

-999 means missing data