import argparse import numpy as np import mitsuba as mi import drjit as dr mi.set_variant("cuda_ad_rgb") import os from PDE3D.BoundaryShape import * from PDE3D.Coefficient import * from PDE3D.utils import * from PDE3D.Solver import * from PDE3D import PATH from python3D.optimization.textures import * from python3D.optimization.sketch import * from python3D.optimization.optimize import * root_directory = os.path.join(PATH, "output3D", "optimizations") def main(): parser = argparse.ArgumentParser(description='''Optimization Sphere''') parser.add_argument('--spp', default = 11, type=int) parser.add_argument('--primalspp', default = 13, type=int) parser.add_argument('--objspp', default = 13, type=int) parser.add_argument('--seedobj', default = 64, type=int) parser.add_argument('--seed', default = 243, type=int) parser.add_argument('--confiter', default = 6, type=int) parser.add_argument('--iternum', default = 512, type=int) parser.add_argument("--lr", default = "0.1", type = float) parser.add_argument("--epsilon", default = "1e-2", type = float) parser.add_argument("--plot", action="store_true") parser.add_argument("--split", default = "normal", type = str) parser.add_argument("--splitdepth", default=250, type=int) parser.add_argument("--computevariance", action = "store_true") parser.add_argument("--scaletexture", default = 1.0, type=float) parser.add_argument("--biastexture", default = 0.0, type=float) parser.add_argument("--conf", default = "1", type = int) parser.add_argument("--verbose", action = "store_true") parser.add_argument("--stepnum", default = 1, type = int) parser.add_argument("--averagepixel", action = "store_true") parser.add_argument("--coeff", default = "source", type = str) parser.add_argument("--screening", default = 0, type = float) parser.add_argument("--measuretime", action = "store_true") parser.add_argument("--constantboundary", action = "store_true") parser.add_argument('--resprimal', default = 16, type=int) parser.add_argument('--restensor', default = 16, type=int) parser.add_argument('--visconf', default = -1, type = int) args = parser.parse_args() res_primal = [args.resprimal,args.resprimal,args.resprimal] res_tensor = [args.restensor,args.restensor,args.restensor] step_num = args.stepnum centered = not args.averagepixel bbox = [[-1,-1],[1, 1]] compute_variance = args.computevariance split_depth = args.splitdepth e_shell = args.epsilon plot = args.plot seed_obj = args.seedobj seed = args.seed spp_obj = 2 ** args.objspp spp = 2 ** args.spp primal_spp = 2 ** args.primalspp conf_per_iter = args.confiter num_iter = args.iternum learning_rate = args.lr bias = args.biastexture scale = args.scaletexture coeff_name = args.coeff screening = args.screening constant_boundary = args.constantboundary vis_conf = [] if args.visconf >= 0: vis_conf = [dr.opaque(mi.UInt32, args.visconf, shape = (1))] if coeff_name == "diffusion" and bias == 0: raise Exception("Please give a positive bias value.") if args.split == "none": split = Split.Naive elif args.split == "agressive": split = Split.Agressive elif args.split == "normal": split = Split.Normal else: raise Exception("No such split is defined.") conf_name = f"conf{args.conf}" centered_name = "centered" if centered else "avg" res_name = f"res{res_primal[0]}" screening_name = f"screen{screening}" spp_name = f"spp{args.primalspp}_{args.spp}" restensor_name = f"restensor{res_tensor[0]}" seed_name = f"seed{seed}" e_name = f"epsilon{args.epsilon}" scale_name = f"scale{scale}-bias{bias}" dirichlet = load_boundary_data(constant_boundary) name = "motorbike-engine" folder_name = os.path.join(PATH, "scenes", name) xml_name = os.path.join(folder_name, "scene.xml") sdf_data = np.load(os.path.join(folder_name, "sdf.npy")) boundary = SDF(sdf_data, dirichlet = dirichlet, scale = 12) obj_name = f"{coeff_name}-{conf_name}" bbox = boundary.bbox bbox_pad = (bbox.max - bbox.min) / 10 bbox_coeff = mi.ScalarBoundingBox3f(bbox.min - bbox_pad, bbox.max + bbox_pad) image_obj = textures[int(args.conf - 1)]() * scale + bias image_begin = np.zeros(res_tensor) #image_begin[int(res_tensor[0] * 3 / 8) : int(res_tensor[0] * 5 / 8), # int(res_tensor[1] * 3 / 8) : int(res_tensor[1] * 5 / 8), # int(res_tensor[2] * 3 / 8) : int(res_tensor[2] * 5 / 8)] = 0.1 image_begin *= scale image_begin += bias if coeff_name == "diffusion": α_obj = TextureCoefficient("diffusion", bbox_coeff, image_obj) α = TextureCoefficient("diffusion", bbox_coeff, image_begin) σ = ConstantCoefficient("screening", screening) data_holder_obj = DataHolder(boundary, α = α_obj, σ = σ) data_holder = DataHolder(boundary, α = α, σ = σ) elif coeff_name == "screening": σ_obj = TextureCoefficient("screening", bbox_coeff, image_obj) σ = TextureCoefficient("screening", bbox_coeff, image_begin) data_holder_obj = DataHolder(boundary, σ = σ_obj) data_holder = DataHolder(boundary, σ = σ) elif coeff_name == "source": f_obj = TextureCoefficient("source", bbox_coeff, image_obj) f = TextureCoefficient("source", bbox_coeff, image_begin) σ = ConstantCoefficient("screening", screening) data_holder_obj = DataHolder(boundary, f = f_obj, σ = σ) data_holder = DataHolder(boundary, f = f, σ = σ) opt_variable_name = f"{coeff_name}.texture.tensor" wos_obj = WosVariable(data_holder_obj) opt_params = [opt_variable_name] wos = WosVariable(data_holder, opt_params = opt_params) input_range = get_range(boundary, bbox_coeff, image_obj) def postprocess(opt, min_val, max_val): opt[opt_variable_name] = dr.clip(opt[opt_variable_name], min_val, max_val) post_process = lambda opt : postprocess(opt, input_range[0] * 0.8, 1.2 * (input_range[1] - input_range[0]) + input_range[0]) def create_path(path): if not os.path.exists(path): os.makedirs(path) folder0 = f"{conf_name}-{coeff_name}" folder0 += "-constDirichlet" if constant_boundary else "" folder1 = f"{res_name}-{scale_name}-{centered_name}" folder1 += f"-{screening_name}" if coeff_name!="screening" else "" folder2 = f"{spp_name}-{seed_name}-{e_name}" folder3 = f"{restensor_name}" folder3 += f"-stepnum{step_num}" if step_num > 1 else "" path_obj = os.path.join(root_directory, "objectives", folder0, folder1) path = os.path.join(root_directory, folder0, folder1, folder2, folder3) print(path) create_path(path_obj) create_path(path) create_path(os.path.join(path, "npy", "primal")) create_path(os.path.join(path, "npy", "grad")) create_path(os.path.join(path, "npy", "tensor")) coeff_obj = wos_obj.input.get_coefficient(coeff_name) plot_coeff(coeff_obj, wos.input.shape, input_range, path_obj, "objective") obj_results = [] for s in range(seed_obj): file = f"{s}.npy" filepath = os.path.join(path_obj, file) if not os.path.isfile(filepath): print(f"Generating objective results for seed {s}.") tensor, std = compute_primals(wos_obj, Split.Agressive, s, bbox, res_primal, spp_obj, centered, split_depth, compute_variance, confs_iter = 16) np.save(filepath, tensor) if compute_variance: filepath_std = os.path.join(path_obj, f"{s}_std.npy") np.save(filepath_std, std) obj_iter = np.load(filepath, allow_pickle = True) obj_results.append(obj_iter) obj_results = np.mean(np.array(obj_results), axis = 0) print("Objective Results are loaded.") wos = optimize(path, wos, wos_obj, obj_results, coeff_name, split, bbox, res_primal, spp, primal_spp, seed, conf_per_iter, split_depth, num_iter, learning_rate, post_process, centered, plot, compute_variance, args.verbose, input_range, measure_time=args.measuretime, vis_set = vis_conf) #if step_num == 1: # wos = optimize_variable(path, wos, wos_obj, obj_results, coeff_name, split, bbox, resolution_primal, spp, primal_spp, seed, conf_per_iter, split_depth, # num_iter, learning_rate, post_process, centered, plot, compute_variance, args.verbose, max_range, out_val = bias, # measure_time=args.measuretime) #else: # for i in range(step_num): # path_iter = os.path.join(path, f"step{i}") # create_path(path_iter) # optimize_variable(path_iter, wos, wos_obj, obj_results, coeff_name, split, bbox, resolution_primal, spp, primal_spp, seed, conf_per_iter, split_depth, # num_iter, learning_rate, post_process, centered, plot, compute_variance, args.verbose, max_range, out_val = bias) #wos.input.upsample2(coeff_name) #wos.get_opt_params(wos.opt_params, opt_params) #print("The input tensor is upsampled, another optimization scheme starts.") if __name__ == "__main__": main()