scripts and helper utilities to simulate Roman grism images

Instructions for setting up NERSC environment needed to run Roman grism simulations

Given NERSC install, hopefully the following is all that is needed for anyone in the roman group

Note, there is now a grizl1 conda environment with updated code installs, but the github links are still in the grizli0 directory

module load conda

conda activate /global/common/software/m4943/grizli1

PYTHONPATH=$PYTHONPATH:/global/common/software/m4943/grizli0/grism_sim/py

export STPSF_PATH="/global/cfs/cdirs/m4943/grismsim/stpsf-data"

export github_dir=/global/common/software/m4943/grizli0/

Full NERSC install (for reference only, most users will need the above option)

module load conda

conda create --prefix /global/common/software/m4943/grizli0 python=3.9

conda activate /global/common/software/m4943/grizli0

conda install numpy scipy astropy

cd /global/common/software/m4943/grizli0

git clone https://github.com/roman-grs-pit/grizli.git

cd grizli

pip install -e .

Following error: "ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. seaborn 0.12.1 requires pandas>=0.25, which is not installed. basemap 1.3.2 requires matplotlib<3.6,>=1.5; python_version >= "3.5", but you have matplotlib 3.9.4 which is incompatible. basemap 1.3.2 requires numpy<1.23,>=1.21; python_version >= "3.7", but you have numpy 1.26.4 which is incompatible."

pip install ".[test]"

pip install stpsf

mkdir /global/cfs/cdirs/m4943/grismsim

export STPSF_PATH="/global/cfs/cdirs/m4943/grismsim/stpsf-data"

cd $STPSF_PATH

cd .. 
#otherwise, get double stpsf-data directory

#check for updated file here https://stpsf.readthedocs.io/en/latest/installation.html
wget https://stsci.box.com/shared/static/kqfolg2bfzqc4mjkgmujo06d3iaymahv.gz

mv kqfolg2bfzqc4mjkgmujo06d3iaymahv.gz  stpsf-data-LATEST.tar.gz

tar xzvf stpsf-data-LATEST.tar.gz

cd /global/common/software/m4943/grizli0

git clone https://github.com/roman-grs-pit/grism_sim.git

git clone https://github.com/roman-grs-pit/star_fields.git

conda install pysynphot (other machines might need pip install pysynphot instead)

PYTHONPATH=$PYTHONPATH:/global/common/software/m4943/grizli0/grism_sim/py

Writing a new sim

New sims can be created by writing a yaml file names sim_config.yaml in a new directory. Said file should have the following format:

stars: path_to_star_cat or null
galaxies: path_to_gal_cat or null
combine_sims: bool
seed: int

# wfi_cen_ra/wfi_cen_dec can be either int/float or dict with start, step, and num
wfi_cen_ra: int, float
wfi_cen_dec:
  start: int, float
  step: int, float
  num: int
wfi_cen_pa:
  start: int, float
  rolls: list
dither: null, int, float or (ra, dec as shown below)
  ra: int, float
  dec: int, float

names_of_sims: list (e.g. [1st_sim, 2nd_sim, …])

1st_sim:
  SCAs: int, list, or “all”
  star_mag_cutoff:               # optional
    brighter_than: int, float    # this is <= mag
    fainter_than: int, float      # this is > mag
  galaxy_mag_cutoff:          # optional
    brighter_than: int, float   # this is <= mag
    fainter_than: int, float     # this is > mag
  key: val - kwargs (e.g. extra_grism_name, extra_ref_name, npsfs, …)
  confdir: str, null
  confs:
    SCA01: str

2nd_sim:
  …

Star catalogs are expected to have the following columns: RA, DEC, magnitude, star_template_index

Galaxy catalogs are expected to have the following columns: RA, DEC, SIM, IDX, [some mag col, whose name is given as an argument]

Galaxy SED templates are saved in "galacticus_FOV_EVERY100_sub_{SIM}.hdf5" Star templates filenames are saved in "star_fields/data/SEDtemplates/input_spectral_STARS.lis"