forked from AmadeusBugProject/artifact_detection
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathRQ3_plot_multi_language.py
185 lines (141 loc) · 7.99 KB
/
RQ3_plot_multi_language.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
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
import numpy as np
import pandas
from matplotlib import pyplot as plt
import matplotlib.patches as mpatches
from artifact_detection_model.utils.Logger import Logger
from datasets.constants import LANGUAGES
from evaluation.stats_utils import evaluate_bootstrap, t_test_x_greater_y, t_test_x_differnt_y, get_box
from file_anchor import root_dir
import seaborn as sns
log = Logger()
OUT_PATH = root_dir() + 'evaluation/out/multi_language/'
CROSS_LANGUAGE_EVALUATION = root_dir() + 'evaluation/out/cross_language/'
language_labels = {
'cpp': 'C++',
'java': 'Java',
'javascript': 'JavaScript',
'php': 'PHP',
'python': 'Python',
}
def main():
bare_stats()
for validation_set_no in ['1', '2']:
cross_project_roc_auc_matrix(validation_set_no)
p_test_single_lang_model_performs_better_than_multi_lang_model(validation_set_no)
p_test_single_lang_model_different_than_multi_lang_model(validation_set_no)
p_test_single_lang_model_different_than_multi_lang_model(validation_set_no)
roc_auc_boxplots(validation_set_no)
multi_model_transferability_table(validation_set_no)
def bare_stats():
multi_df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
for lang in LANGUAGES:
columns = ['roc-auc_' + lang + '_researcher_' + x for x in ['1', '2']]
multi_df[columns].describe().to_csv(OUT_PATH + lang + '_performance.csv')
def p_test_single_lang_model_performs_better_than_multi_lang_model(validation_set_no):
multi_df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
rep_df = pandas.DataFrame()
for lang in LANGUAGES:
df = pandas.read_csv(CROSS_LANGUAGE_EVALUATION + lang + '_artifact_detection_cross_language_resample_summary.csv')
rep = t_test_x_greater_y(df['roc-auc_' + lang + '_researcher_' + validation_set_no],
multi_df['roc-auc_' + lang + '_researcher_' + validation_set_no],
lang, 'multilang') # one sided, x greater y
rep['model'] = 'multilang'
rep_df = rep_df.append(rep)
rep_df.to_csv(OUT_PATH + 'single_lang_model_better_on_its_own_language_than_multilang_model_VS' + validation_set_no + '.csv')
def p_test_single_lang_model_different_than_multi_lang_model(validation_set_no):
multi_df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
rep_df = pandas.DataFrame()
for lang in LANGUAGES:
df = pandas.read_csv(CROSS_LANGUAGE_EVALUATION + lang + '_artifact_detection_cross_language_resample_summary.csv')
rep = t_test_x_differnt_y(df['roc-auc_' + lang + '_researcher_' + validation_set_no],
multi_df['roc-auc_' + lang + '_researcher_' + validation_set_no],
lang, 'multilang') # one sided, x greater y
rep['model'] = 'multilang'
rep_df = rep_df.append(rep)
rep_df.to_csv(OUT_PATH + 'single_lang_model_than_multilang_model_VS' + validation_set_no + '.csv')
def cross_project_roc_auc_matrix(validation_set_no):
columns = ['roc-auc_' + x + '_researcher_' + validation_set_no for x in LANGUAGES]
cm = []
for lang in LANGUAGES:
df = pandas.read_csv(CROSS_LANGUAGE_EVALUATION + lang + '_artifact_detection_cross_language_resample_summary.csv')
df = df[columns].mean()
cm.append(df.to_list())
df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
df = df[columns].mean()
cm.append(df.to_list())
fig, ax = plt.subplots() #figsize=(3, 3)
sns.heatmap(cm,
ax=ax,
# linewidths=0.01,
# linecolor='k',
cmap="viridis",
annot=True,
annot_kws={'fontsize':'large'},
xticklabels=[language_labels[x] for x in LANGUAGES],
yticklabels=[language_labels[x] for x in LANGUAGES] + ['Multi language'])
plt.yticks(rotation=0)
ax.set(ylabel="Model language", xlabel='Validation set ' + validation_set_no + ' language', title='ROC-AUC')
plt.tight_layout()
plt.savefig(OUT_PATH + 'multi_language_project_roc_auc_matrix_VS' + validation_set_no + '.pdf')
def roc_auc_boxplots(validation_set_no):
multi_df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
results_df = pandas.DataFrame()
multi_boxes = []
lang_boxes = []
for lang in LANGUAGES:
lang_df = pandas.read_csv(CROSS_LANGUAGE_EVALUATION + lang + '_artifact_detection_cross_language_resample_summary.csv')
mult_res = evaluate_bootstrap(multi_df['roc-auc_' + lang + '_researcher_' + validation_set_no], 'Multilang' + '_VS' + validation_set_no)
lang_res = evaluate_bootstrap(lang_df['roc-auc_' + lang + '_researcher_' + validation_set_no], lang + '_VS' + validation_set_no)
results_df = results_df.append(pandas.DataFrame([mult_res]))
results_df = results_df.append(pandas.DataFrame([lang_res]))
multi_boxes.append(get_box(mult_res))
lang_boxes.append(get_box(lang_res))
results_df.to_csv(OUT_PATH + 'multi_language_roc_auc_bootstrap_VS' + validation_set_no + '.csv')
fig, ax1 = plt.subplots(figsize=(8, 4))
space = 0.2 # boxprops=dict(facecolor='tab:blue'),
boxplot1 = ax1.bxp(multi_boxes, showfliers=False, widths=0.4, patch_artist=True, boxprops=dict(facecolor='lightcyan'), medianprops=dict(color="black", linewidth=1.5), positions=np.arange(5)-space,)
ax2 = ax1.twinx()
boxplot2 = ax2.bxp(lang_boxes, showfliers=False, widths=0.4, patch_artist=True, boxprops=dict(facecolor='lightgreen'), medianprops=dict(color="black", linewidth=1.5), positions=np.arange(5)+space,)
ax1_lim = ax1.get_ylim()
ax2_lim = ax2.get_ylim()
ax1.set_ylim(0.90, 0.97)
ax2.set_ylim(0.90, 0.97)
ax2.set_yticks([])
ax1.set_xticks(np.arange(5))
ax1.set_xticklabels([f'{label}' for label in language_labels.values()])
ax1.set_ylabel('ROC-AUC')
ax1.set_title('')
plt.sca(ax1)
plt.legend(handles=[mpatches.Patch(color='lightcyan', label='Multi language model'),
mpatches.Patch(color='lightgreen', label='Language specific model')],
loc='lower left')
plt.tight_layout()
plt.savefig(OUT_PATH + 'multi_language_roc_auc_boxplots_VS' + validation_set_no + '.pdf')
def multi_model_transferability_table(validation_set_no):
comb_roc_auc = pandas.DataFrame(columns=[language_labels[l] for l in LANGUAGES] + ['Multi language'])
for lang in LANGUAGES:
df = pandas.read_csv(CROSS_LANGUAGE_EVALUATION + lang + '_artifact_detection_cross_language_resample_summary.csv')
roc_auc = []
for l in LANGUAGES:
roc_auc.extend(df['roc-auc_' + l + '_researcher_' + validation_set_no].to_list())
comb_roc_auc[language_labels[lang]] = roc_auc
df = pandas.read_csv(OUT_PATH + 'artifact_detection_multi_language_model_resample_summary.csv')
roc_auc = []
for l in LANGUAGES:
roc_auc.extend(df['roc-auc_' + l + '_researcher_' + validation_set_no].to_list())
comb_roc_auc['Multi language'] = roc_auc
rep_df = pandas.DataFrame()
for column in comb_roc_auc.columns:
if column == 'Multi language':
continue
rep = t_test_x_greater_y(comb_roc_auc['Multi language'],
comb_roc_auc[column],
'Multi language', column)
rep_df = rep_df.append(rep)
rep_df.to_csv(OUT_PATH + 'multilang_model_better_transfer_than_single_lang_model_transfer_VS' + validation_set_no + '.csv')
rep_df = rep_df[rep_df['test'] == 'wilcoxon']
comb_roc_auc = comb_roc_auc.mean()
comb_roc_auc.to_csv(OUT_PATH + 'transferability_mean_over_all_language_performance_VS'+ validation_set_no + '.csv')
comb_roc_auc.T.to_latex(OUT_PATH + 'transferability_mean_over_all_language_performance_VS' + validation_set_no + '.tex', float_format="%.2f")
if __name__ == "__main__":
main()