import argparse import multiprocessing as mp from functools import partial import numpy as np from acoustics_2d_interface_maze import setup as maze_setup from acoustics_2d_interface_random_medium import setup as random_setup # Time/steps/samples steps_map = { "continuous": (2.0, 101, 2000), "discontinuous": (2.0, 101, 2000), "inclusions": (2.0, 101, 4000), "maze": (4.0, 201, 2000), } def mp_wrapper( seed, discontinuous, inclusions, maze, output_dir="/mnt/home/polymathic/ceph/the_well/testing_before_adding/clawpack_data/acoustics_2d_variable/", ): if discontinuous: run_func = partial( inner_gen_sample, discontinuous=True, inclusions=False, maze=False, output_dir=output_dir, ) num_samples = 2000 elif inclusions: run_func = partial( inner_gen_sample, discontinuous=False, inclusions=True, maze=False, output_dir=output_dir, ) num_samples = 4000 elif maze: run_func = partial( inner_gen_sample, discontinuous=False, inclusions=False, maze=True, output_dir=output_dir, ) num_samples = 2000 else: run_func = partial( inner_gen_sample, discontinuous=False, inclusions=False, maze=False, output_dir=output_dir, ) num_samples = 2000 cores = mp.cpu_count() seeds = seed.spawn(num_samples) with mp.Pool(cores // 2) as pool: pool.map(run_func, seeds) # run_func(seeds[0]) def inner_gen_sample( seed=0, discontinuous=False, inclusions=False, maze=False, output_dir="" ): """ Iterate num samples times and enerate sample file. Use it to overwrite qinit, then run .make output to generate trajectory. Make sure overwrite is False in the make file before running. """ # Check conditions and set up names file_suffix = f"{str(seed.bit_generator.seed_seq.entropy)}_{str(seed.bit_generator.seed_seq.spawn_key)}" if discontinuous: file_suffix = "discontinuous_" + file_suffix run_func = partial( random_setup, seed=seed, include_splits=True, include_inclusions=False, outdir=output_dir + file_suffix, T_max=2.0, num_steps=101, ) elif inclusions: file_suffix = "inclusions_" + file_suffix run_func = partial( random_setup, seed=seed, include_splits=True, include_inclusions=True, outdir=output_dir + file_suffix, T_max=2.0, num_steps=101, ) elif maze: file_suffix = "maze_" + file_suffix run_func = partial( maze_setup, seed=seed, outdir=output_dir + file_suffix, T_max=4.0, num_steps=201, ) else: file_suffix = "continuous_" + file_suffix run_func = partial( random_setup, seed=seed, include_splits=False, include_inclusions=False, outdir=output_dir + file_suffix, T_max=2.0, num_steps=101, ) claw = run_func(output_dir + file_suffix) claw.run() if __name__ == "__main__": # print(len(gases)) parser = argparse.ArgumentParser( description="Generate initial conditions for 2D Euler quadrants" ) # parser.add_argument('--num_samples', type=int, default=1000, help='Number of samples to generate') parser.add_argument( "--discontinuity", action="store_true", help="Whether to generate random samples", ) parser.add_argument( "--inclusions", action="store_true", help="Whether to generate random samples" ) parser.add_argument( "--switch_to_maze", action="store_true", help="Whether to generate random samples", ) # parser.add_argument('--bc', type=str, default='extrap', help='Boundary conditions') # parser.add_argument('--gas_index', type=int, default=0, help='Index of gas to use (0-9 inclusive)') parser.add_argument( "--seed", type=int, default=0, help="Seed for random samples - use different one per gas/bc if par", ) parser.add_argument( "--raw_output_dir", type=str, default=" /mnt/home/polymathic/ceph/the_well/testing_before_adding/clawpack_data/", help="Directory to store raw output", ) args = parser.parse_args() seed = np.random.default_rng( args.seed + 100 * int(args.discontinuity) + 1000 * int(args.inclusions) + 10000 * int(args.switch_to_maze) ) mp_wrapper(seed, args.discontinuity, args.inclusions, args.switch_to_maze)