Test ground for long-term simulation generation
Sims cloned at /sdf/data/rubin/shared/fbs_sims/sims_longterm_tests/
:
- Baseline simulation - copy of baseline from v4.3.1; in
baseline
folder. - Weather simulations - copy of baseline from v4.3.1; in
weather
folder. Note:had to move what seems to be a buggy file in a subfolderbuggy
. - Plots - vector metric Nvisits (Nvisits per filter + all filters) over time, for each simulation.
Goal is to have two kinds of simulations:
- Chimera simuations - a weather simulation up to date X, then baseline simulation after date X.
- Continuing/bespoke simulations - a weather simulation up to date X, then continue simulation using baseline configuration.
Code structure (all in scripts
folder in this repo):
- primary script is
run.py
with various options. usesconfig.yml
and is easily run byrun.sh
(see the script for how to enable various stages of the analysis).run.py
calls three helper functions, inget_fonvtime.py
,get_chimera.py
, andget_bespoke.py
As an example, results for two cutoff_dates
(X above) are shown in notebooks/dev-analysis.ipynb
.
Misc notes:
- bespoke sims take a while (9-14hrs), hence the need for in-queue scripts for each weather sim; see more in
run.sh
. - everything except bespoke sims generations can happen on an interactive node - but for ease,
run.sh
includes the option to run various stages in queue as well.