-
Notifications
You must be signed in to change notification settings - Fork 0
/
plot_csv.py
161 lines (130 loc) · 4.95 KB
/
plot_csv.py
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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
import os
import glob
import numpy as np
from matplotlib import pyplot as plt
from typing import List
from attack_types import run_names
WANDB_DIR = "./wandb/"
COLORS = ['#332288', '#117733', '#44AA99', '#88CCEE', '#DDCC77', '#CC6677', '#D879D4', '#AA4499', '#882255',
'#37001C', '#A26F85']
FID_VALUES = [5.52843996652564, 13.5188010775728, 14.4754941578319, 13.0412494709869, 10.5513145035911,
15.0473063432149, 14.1113076272244, 10.6831156682874, 4.60950168091205, 6.7844135449148]
def read_csv_data(dir: str) -> [np.ndarray, List[str]]:
"""
Read csv data into a numpy array and a list.
:return: data numpy array and label list
"""
old_pwd = os.getcwd()
os.chdir(WANDB_DIR + dir)
files = glob.glob("*")
files.sort(key=os.path.getmtime)
data, label = [], []
for file in files:
filename = os.fsdecode(file)
# if filename not in ['synonym-word', 'typo-char', 'naive-char', 'original-control']:
# continue
label.append(filename)
with open(filename) as f:
data.append([float(v[1:-1]) for v in f.read().splitlines()[1:]])
os.chdir(old_pwd)
return np.asarray(data), label
def plot_histogram(data: np.ndarray, label: List[str], title: str, directory: str, y_max: float):
"""
Plot data in a histogram and save the plot.
:param data: data that gets plotted
:param label: label for the data
:param title: title for the whole plot
:param directory: dictionary where to save all histograms
"""
padding = (np.max(data) - np.min(data)) * 0.1
x_min = np.min(data) - padding
x_max = np.max(data) + padding
bins = np.linspace(np.min(data), np.max(data), num=10, endpoint=False)
plt.hist(data.T, bins, color=COLORS, histtype='bar', label=label)
plt.legend(prop={'size': 10})
plt.title(title)
file_name = title.replace(" ", "_").lower()
plt.savefig('./wandb/plots/' + file_name + '.pdf')
plt.show()
plt.close()
fig, axs = plt.subplots(5, 2, figsize=(10, 16))
y, x = 0, 0
for i in range(len(label)-1):
if x == 0 and y == 0:
d = np.vstack((data[i], data[len(label)-1])).T
l = [label[i], label[-1]]
c = [COLORS[i], COLORS[-1]]
else:
d = data[i].T
l = label[i]
c = COLORS[i]
axs[y, x].hist(d, bins, color=c, histtype='bar', label=l, rwidth=0.8)
axs[y, x].legend(prop={'size': 12})
axs[y, x].set_ylim([0, y_max])
axs[y, x].set_xlim([x_min, x_max])
axs[y, x].set_xlabel("Cosine Similarity")
axs[y, x].set_ylabel("Count")
if x == 0:
x += 1
else:
y += 1
x = 0
fig.suptitle(title, fontsize=14)
plt.tight_layout()
plt.savefig('./wandb/plots/' + directory + '.pdf')
plt.show()
plt.close()
def plot_chart(data: np.ndarray, label: List[str], title: str):
"""
Plot mean data in a chart and save the plot.
:param data: data that gets plotted
:param label: label for the data
:param title: title for the whole plot
"""
plt.figure(constrained_layout=True)
y_pos = np.arange(len(label))
means = [np.mean(x) for x in data]
plt.barh(y_pos, means, align='center', color=COLORS)
plt.yticks(y_pos, labels=label)
plt.gca().invert_yaxis()
plt.xlabel("Cosine Similarity")
plt.title(title)
padding = (max(means) - min(means)) * 0.1
x_min = min(means) - padding
x_max = max(means) + padding
plt.xlim([x_min, x_max])
file_name = title.replace(" ", "_").lower()
plt.savefig('./wandb/plots/' + file_name + '.pdf')
plt.show()
plt.close()
def plot_fid_chart(title: str):
plt.figure(constrained_layout=True)
labels = run_names + ["random"]
y_pos = np.arange(len(labels))
means = FID_VALUES + [FID_VALUES[0]]
plt.barh(y_pos, means, align='center', color=COLORS)
plt.yticks(y_pos, labels=labels)
plt.gca().invert_yaxis()
plt.xlabel("Cosine Similarity")
plt.title(title)
padding = (max(means) - min(means)) * 0.1
x_min = min(means) - padding
x_max = max(means) + padding
plt.xlim([x_min, x_max])
file_name = title.replace(" ", "_").lower()
plt.savefig('./wandb/plots/' + file_name + '.pdf')
plt.show()
plt.close()
def main():
data, label = read_csv_data("csv-files-cosine-sim/")
plot_histogram(data, label, "Image Cosine Similarity", "image_cosine_sim", 5000)
plot_chart(data, label, "Mean Image Cosine Similarity")
data, label = read_csv_data("csv-files-image-text-sim/")
plot_histogram(data, label, "Image Text Similarity", "image_text_sim", 6000)
plot_chart(data, label, "Mean Image Text Similarity")
data, label = read_csv_data("csv-files-image-caption-sim/")
plot_histogram(data, label, "Image Caption Similarity", "image_caption_sim", 5200)
plot_chart(data, label, "Mean Image Caption Similarity")
plot_fid_chart("Clean FID Score")
if __name__ == '__main__':
main()