NNblendproject/plot_functions.py at master · paxtonsc/NNblendproject · 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
125
126
127
128
129
import torch
import matplotlib.pyplot as plt
import numpy as np
import os

def plot_data_distribution(X):
	#X = torch.flatten(X)
	X = X.reshape(X.shape[0]*5, int(X.shape[1]/5))

	fig, axs = plt.subplots(3, 3)
	for i in range(9):
		axs[int(i/3), i%3].hist(X[:,i], bins=20, alpha=0.4)

	axs[0, 0].set_title('x-pos from center')
	axs[0, 1].set_title('y-pos from center')
	axs[0, 2].set_title('gal std')
	axs[1, 0].set_title('e1')
	axs[1, 1].set_title('e2')
	axs[1, 2].set_title('noise')
	axs[2, 0].set_title('gal flux')
	axs[2, 1].set_title('psf flux')
	axs[2, 2].set_title('psf std')

	for ax in axs.flat:
		ax.set(xlabel = 'data value', ylabel = '# occurances')
		#ax.label_outer()

	fig.tight_layout(pad=0.3)


	fig_name = os.path.join('charts', 'data_distribution.png')
	plt.savefig(fig_name)

def plot_loss_acc(acc_test, acc_train, acc_1=None, acc_2=None, acc_3=None, acc_4=None, acc_5=None):
	"""
	plots accuracy after the nn is done training
	saves in 'charts' folder

	:param acc
	:param acc_train
	"""

	if acc_3 and acc_2:
		plt.plot(acc_1, label='test accuracy 1')
		plt.plot(acc_2, label='test accuracy 2')
		plt.plot(acc_3, label='test accuracy 3')
		plt.plot(acc_4, label='test accuracy 4')
		plt.plot(acc_5, label='test accuracy 5')
	plt.plot(acc_test,label='test accuracy')
	plt.plot(acc_train, label='train accuracy')
	plt.xlabel('epochs (in hundreds)')
	plt.legend()
	file_name = os.path.join('charts', 'nn_accuracy.png')
	plt.savefig(file_name)


def NN_vs_source(y_source, y_NN, GALS=5, PARAM_PER_GAL=9):
	class_array = np.zeros((GALS+1, GALS+1))
	for i in range(len(y_source)):
		index1 = torch.where(y_source[i] == 1)[0]
		index2 = torch.where(y_NN[i] == 1)[0]
		class_array[index1+1][index2+1] += 1

	class_array = (class_array.T/class_array.sum(axis=1)).T

	plt.matshow(class_array, cmap='YlOrRd')
	plt.ylabel('# identified by SE')
	plt.xlabel('# predicted by NN')
	plt.suptitle('NN predictions of SE vs SE')
	for (i, j), z in np.ndenumerate(class_array):
		plt.text(j, i, '{:0.3f}'.format(z), ha='center', va='center')
	file_name = os.path.join('charts', 'nn_vs_se.png')
	plt.savefig(file_name)


def plot_class_distribution(y_vals, x_vals, NN=False, GALS=5, PARAM_PER_GAL=9):
	"""
	create matrix viz of galsim num generated galaxies
	to number recgnoized by source extractor or number that nn thinks
	source extractor will recognize.

	:param y_vals: true number predicted by SE
	:param x_vals: true parameters from GalSim
	"""

	matrix = np.zeros((6,6))
	num_drawn = np.zeros(GALS)
	j = 0
	for x in x_vals:
		num = 0
		for i in range(5):
			if x[PARAM_PER_GAL*i+1] or x[PARAM_PER_GAL*i]:
				num += 1
			else:
				break
		num_drawn[num -1] += 1
		index = torch.where(y_vals[j] == 1)[0]
		matrix[num][index+1] += 1
		j+=1
		assert(index+1 >0)

	matrix = (matrix.T/matrix.sum(axis=1)).T

	#sums = torch.sum(y_vals, dim=0)
	#gals = ['one', 'two', 'three', 'four', 'five']
	#plt.bar(gals, num_drawn, alpha=0.6, label='gal sim galaxies drawn')
	#plt.bar(gals, sums, alpha=0.6, label='Source extractor galaxies found')
	#plt.xlabel('number of galaxies')
	#plt.ylabel('number of ocurances in training set')
	#plt.suptitle('Training Data Makeup')
	#if (NN):
	#	plt.suptitle('NN predictions')
	#plt.legend()
	plt.matshow(matrix, cmap='YlOrRd')
	plt.xlabel('# identified by SE')
	plt.ylabel('# drawn by GalSim')
	plt.suptitle('Training Data Makeup')
	if (NN):
		plt.suptitle('NN predictions')
	for (i, j), z in np.ndenumerate(matrix):
		plt.text(j, i, '{:0.3f}'.format(z), ha='center', va='center')

	path = os.path.join('charts', 'data_makeup.png')
	if NN:
		path = os.path.join('charts', 'data_makeup.NN.png')

	plt.savefig(path)