ms_thesis/experimental_analysis.py at main · probabilis/ms_thesis · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
import numpy as np
import torch
import matplotlib.pyplot as plt
import pandas as pd

from utils.env_utils import PATHS, read_sim_dat_from_csv, plotting_style, main_colormap
from exp_data_processing.read import read_csv
from params.opt_params import labyrinth_data_params, exp_data_params, get_DataParameters
from exp_data_processing.postprocessing import quality_scores

from spectrum_analysis import radial_wavelength_spectrum
from exchange_length import reduced_exp_image_width


if __name__ == "__main__":


    num_iters = 5_000
    dataset = "data_01"

    file_indices = [1, 2, 3, 4, 5, 6, 7]
    arrangement = [3, 2, 1, 5, 4, None, 6]
    FILE_AMOUNT = 6

    domain_pattern_lengths = np.zeros(FILE_AMOUNT)

    INPUT_PATH = PATHS.BASE_EXPDATA

    # ---------------------------------------------------------------

    self_picked_indices = [14, 11, 11, 7, 4, 1]

    fig, axs = plt.subplots( 5, 6, figsize = (16,10))


    for ii, OPT_FILE in enumerate( file_indices ):

        recording = f"00{OPT_FILE}"

        INPUT_FILE_PATH = INPUT_PATH / f"{dataset}/csv/mcd_slice_{recording}.csv"
        OUTPUT_PATH = PATHS.BASE_EXPDATA / dataset / "opt" / recording

        df = pd.read_csv(OUTPUT_PATH / f"ranking.csv", index_col = 0)

        cut = df["perimeter"].quantile(0.6)
        df_filt = df[df["perimeter"] <= cut].copy()

        best = df_filt.sort_values(["fisherJ", "perimeter"], ascending = [False, True])


        _gamma = best.iloc[0]["gamma"]
        _lambda = best.iloc[0]["lambda"]

        if ii == 6:
            _gamma = df.iloc[1]["gamma"]
            _lambda = df.iloc[1]["lambda"]

        u_exp = read_csv(INPUT_FILE_PATH, PLOT = False)
        df_energies, u_sim = read_sim_dat_from_csv(OUTPUT_PATH, exp_data_params.N, num_iters, _gamma, exp_data_params.epsilon, _lambda)

        if arrangement[ii] is not None:

            kk = arrangement[ii] - 1
            self_picked_new_index = self_picked_indices[kk]
            self_picked_gamma = df.iloc[self_picked_new_index]["gamma"]
            self_picked_lambda = df.iloc[self_picked_new_index]["lambda"]
            self_df_energies, self_u_sim = read_sim_dat_from_csv(OUTPUT_PATH, exp_data_params.N, num_iters, self_picked_gamma, exp_data_params.epsilon, self_picked_lambda)

            new_index = arrangement[ii] - 1

            axs[0, new_index].imshow(u_exp, cmap = main_colormap, origin="lower", extent=(0,1,0,1))


            u_sim = torch.tensor(u_sim.values)
            axs[1, new_index].set_title(f"$\\gamma$ = {_gamma} \n $\\lambda$ = {_lambda}")
            axs[1, new_index].imshow(u_sim, cmap = main_colormap, origin="lower", extent=(0,1,0,1))


            self_u_sim = torch.tensor(self_u_sim.values)
            axs[2, new_index].imshow(self_u_sim, cmap = main_colormap, origin="lower", extent=(0,1,0,1))
            axs[2, new_index].set_title(f"$\\gamma$ = {self_picked_gamma} \n $\\lambda$ = {self_picked_lambda}")


            fisherJ, perimeter, W = quality_scores(u_exp, self_u_sim)
            W = torch.where(torch.abs(self_u_sim) > 0.15, 0, self_u_sim)
            axs[3, new_index].imshow(W.float(), cmap = main_colormap, origin="lower", extent=(0,1,0,1) )
            axs[3, new_index].set_title(f"$S(u)$ = {fisherJ:.4g} \n $P(u)$ = {perimeter:.5g}")


            results = radial_wavelength_spectrum(self_u_sim, exp_data_params.gridsize/exp_data_params.N, plot = False)
            axs[4, new_index].loglog(results["k"].cpu(), results["profile"].cpu(), lw=2, color = "blue", label = "opt.")
            axs[4, new_index].axvline(results["k_peak"], linestyle="--", label=f"$k^*$ = {results["k_peak"]:.4g}", color = "blue")


            results_raw = radial_wavelength_spectrum(u_exp, exp_data_params.gridsize/exp_data_params.N, plot = False)
            axs[4, new_index].loglog(results_raw["k"].cpu(), results_raw["profile"].cpu(), lw=2, color = "red", label = "raw")
            axs[4, new_index].axvline(results_raw["k_peak"], linestyle="--", label=f"$k^*$ = {results_raw["k_peak"]:.4g}", color = "red")
            domain_pattern_length = results["wavelength_peak"]/2 * reduced_exp_image_width * 1e6
            axs[4, new_index].set_title(f"$D_p = $ {domain_pattern_length:.4g}")
            axs[4, new_index].legend()


            for jj in range(3):
                axs[jj, new_index].set_box_aspect(1)
                axs[jj, new_index].axes.get_xaxis().set_ticks([])
                axs[jj, new_index].axes.get_yaxis().set_ticks([])


            domain_pattern_lengths[new_index] = domain_pattern_length
            kk += 1

    print(domain_pattern_lengths)


    fig.tight_layout()
    plt.show()