Spherical harmonic sampler for Hydra
All actual parameters are set in the python script directly / hard coded (sorry Phil!)
Remember to update all paths.
COMMAND LINE ARGS: -dir (or --directory) sets the output directory for the precomputation and results ONLY, the slurm output (which will tell you when precomputation is done, any errors and if it succeeds also writes what your output folder is to easier link the two)
-data_seed (or --data_seed) sets the random seed for the precomputation part, i.e. also the noise on the data.
-jobid (or --jobid) parsed the array-job id to the script, the radnom seed for the specific sample is set by 100 * jobid + sample_number
OUTPUT: The data is saved in .npz files as results_{data_seed}_{random_seed}.npz where random_seed refers to the specific samples random seed
The precomputations are saved last because all the times for the run are saved in this file too. The precomputation is saved as precomputation_{data_seed}_{jobid} since one precomputation covers the entire job.