System packages (Fedora): R
R packages: Rcpp dplyr RColorBrewer ggplot2 tidyr gridExtra stringr
./generate_data.sh $multiplier $count
Will generate a $count sets of data to be graphed, with sizes (n,p, etc.) scaled by $multiplier. Graph scripts assume the multiplier is either 10 or 1. Output is stored in ./simulated_data.
./generate_lethal_data.sh $multiplier $count
Will do the same, but will generate data with lethal pairs. For the moment this is only used in one graph. Output is stored in ./simulated_data.
./fit_all_1204.sh $threads
Will check whether each generated dataset has already been (fitted/checked?) by xyz, and, if it hasn't, do so now. This will be done on $threads separate threads (2 by default).
./generate_graphs.sh [large/small/store]
Will (re)create every graph. 'large' will use n=10,000, p=1,000. 'small' will use n=1,000, p=100. 'store' will move every graph currently present into a subdirectory "pdfs_date +%s" (e.g. pdfs_1548386299). To find out when such a directory was made (in a more readable fashion), use date --date=@1548386299 (adjusting the date as required).
N.B. attempts using different values of L (number of projections, for l_diff) need to be run manually (using ./fits_xyz_only $simulated_data/filename $L write).