various bookkeeping procedures and analysis routines for Non-Parametric Schwarzschild Modeling (NPSM)
For use with the NPSM code described in Jardel et al (2013 ApJ 763 91). This repo is divided into 3 sections:
- TACC - routines to be run on the Texas Advanced Computing Center (Lonestar). These include routines to manage job submission, parameter space sampling, chi^2 analysis, and interface with PSQL database storing results
- local - routines to be run on your local machine. These are mostly designed to perform analysis on results obtained on TACC and stored in the database (Corral)
- R - routines for interactive visualizations.
DESCRIPTIONS:
- TACC
plotres.py - Main analysis routine to determine dark matter densities and corresponding uncertainties. Computes them from sliding biweight of k-lowest chi^2 points per bin for marginalized chi^2 curves. Makes Figure 6 in Jardel et al. (2013 ApJ 763, 91)
plotres_multi-param.py - Secondary analysis routine to determine dark matter densities and their uncertainties. Computes the joint 1-sigma confidence band (in contrast to the point-wise band determined by plotres.py). Also does a better job at interpolating between radial bins. Makes gray band from Figure 1 in Jardel & Gebhardt (ApJL 775L, 30 ).
rungrid.py - Sets up a multi-dimensional grid of parameter combinations to run, a brute force method. Usually start off modeling with rungrid, then do something more sophisticated like smartgrid.py. Contains Launcher class for submitting jobs to Lonestar.
smartgrid.py - Uses iterative refinement technique described in Jardel et al. (2013 ApJ 763, 91) to sample parameter space. Finds areas of low chi^2 and takes fractional steps in each direction. Can be automated to run with crontab using rsubmit.s and submit.s.
submit.s - Script to run smartgrid.py
rsubmit.s - Driver script to call submit.s. Place a call to rsubmit.s inside of a crontab.
update_db.py - Inserts results from a model into the PostgreSQL database on Corral. Place a local copy in base/.
runall.s - Main driver script to run NPSM.
mplot.py - Makes individual M(r) and circular speed plots for a single dSph. Also plots location of the Wolf et al. (2010) mass estimator for comparison to the full profile.
mk_mplot4p.py - Makes a 4-panel grid of M(r) plots.
run_master_mplot.csh - Shell script to do the legwork necessary to run mplot.py and mk_mplot4p.py.
- local
dsphdm2.py - Computes the dark matter density profile by subtracting the stellar profile from the total density profile (determined from plotres and plotres_multi on TACC). Relies on SSP models from Maraston et al. (2005) in SSP.tab. Writes out *.res files for reading with npsm_common.py. Also fits a power law with Monte Carlo methods (deprecated, use calc_est.py instead).
npsm_common.py - Holds Galaxy class. A Galaxy instance consists of a set of
matplotlib axes with the full dark matter density profile
ready for plotting in a combined plot with plotall.py.
First run dsphdm2.py to calculate the densities.
plotall.py - Constructs a multi-panel plot for a number of galaxies at the same time. Creates plots like Figure 1 in Jardel & Gebhardt 2013 (ApJL 775L, 30 ),
plotall_single_panel.py - Puts all the Galaxy instances on a single panel like Figure 2 of Jardel & Gebhardt (2013 ApJL, 775L, 30 ). Calls plot_dmprof_all() method of Galaxy.
plot_monents_univ.py - Makes plots of Gauss-Hermite moments (V, sigma, h3, h4) for any dSph. Makes something like Figure 4 of Jardel et al. (2013, 763, 91).
get_res_all.s - Script to grab all the analysis data (chiXXX.out) from Lonestar, then run dsphdm2.py on each dSph. Puts .res files in right location for a run with plotall.py.
cmdplot.py - Gets stellar population parameters by fitting isochrones to color-magnitude diagrams. Returns [Fe/H] and t_age.
chiPlotter.py - Plays around with results interactively. Lets the user specify a delta chi^2 threshold and see how the profile changes. Can also show where (spatially) most of the contribution to total chi^2 comes from. Writes to file readable by the Shiny app in R/.
- R
ui.R server.R
Routines to run Shiny App to explore output of chiPlotter.py. See http://rstudio.github.io/shiny/tutorial/ for details on how to use Shiny.