-
Notifications
You must be signed in to change notification settings - Fork 9
/
visualize.py
281 lines (253 loc) · 15.9 KB
/
visualize.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
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
from curses import BUTTON1_TRIPLE_CLICKED
import os
from os.path import join
import json
import torch
import networkx as nx
import numpy as np
import matplotlib.pyplot as plt
# import pygraphviz as pgv
from matplotlib.collections import PolyCollection
from matplotlib import cm
from scipy.stats import poisson
from sklearn.decomposition import PCA
from mpl_toolkits.mplot3d import Axes3D
from sco_models.visualization import nodes_edges_correlation
from sco_models.visualization import nodes_edges_compressed_graph_correlation
from sco_models.model_hgt import HGTVulGraphClassifier
def polygon_under_graph(x, y):
"""
Construct the vertex list which defines the polygon filling the space under
the (x, y) line graph. This assumes x is in ascending order.
"""
# print([(x[0], 0.), *zip(x, y), (x[-1], 0.)])
return [(x[0], 0.), *zip(x, y), (x[-1], 0.)]
def plot_graph(nxg):
model = HGTVulGraphClassifier(nxg, node_feature='nodetype', hidden_size=128, num_layers=2,num_heads=8, use_norm=True, device='cpu')
graph = model.symmetrical_global_graph
ag = pgv.AGraph(strict=True, directed=False)
for u, v, k in graph.canonical_etypes:
ag.add_edge(u, v, label=k)
ag.layout('dot')
ag.draw('graph.png')
if __name__ == '__main__':
## Compressed graph forensic
bug_type = {'access_control': 57, 'arithmetic': 60, 'denial_of_service': 46,
'front_running': 44, 'reentrancy': 71, 'time_manipulation': 50,
'unchecked_low_level_calls': 95}
# bug_type = {'access_control': 57}
# nxg = './experiments/ge-sc-data/source_code/access_control/clean_57_buggy_curated_0/cfg_cg_compressed_graphs.gpickle'
# plot_graph(nxg)
# creation_path = '/home/minhnn/minhnn/ICSE/EtherSolve_ICPC2021_ReplicationPackage/data/reentrancy/creation/'
# output = '/home/minhnn/minhnn/ICSE/ge-sc/forensics'
# runtime_path = '/home/minhnn/minhnn/ICSE/EtherSolve_ICPC2021_ReplicationPackage/data/reentrancy/runtime/'
# creation_files = [join(creation_path, f) for f in os.listdir(creation_path) if f.endswith('.sol')]
# runtime_files = [join(runtime_path, f) for f in os.listdir(runtime_path) if f.endswith('.sol')]
# nodes_edges_correlation([creation_files], join(output, 'creation_node_edge_correlation.png'))
# nodes_edges_correlation([runtime_files], join(output, 'runtime_node_edge_correlation.png'))
# for bug, count in bug_type.items():
# fig, axs = plt.subplots(2, 3, figsize=(10,10), sharex=True, sharey=True)
# compressed_graph = f'./experiments/ge-sc-data/source_code/{bug}/clean_{count}_buggy_curated_0/cfg_cg_compressed_graphs.gpickle'
# nodes_edges_compressed_graph_correlation(compressed_graph, axs[0, 0], title='CFG + CG')
# compressed_graph = f'./experiments/ge-sc-data/source_code/{bug}/clean_{count}_buggy_curated_0/cfg_compressed_graphs.gpickle'
# nodes_edges_compressed_graph_correlation(compressed_graph, axs[0, 1], title='CFG')
# compressed_graph = f'./experiments/ge-sc-data/source_code/{bug}/clean_{count}_buggy_curated_0/cg_compressed_graphs.gpickle'
# nodes_edges_compressed_graph_correlation(compressed_graph, axs[0, 2], title='CG')
# compressed_graph = f'./experiments/ge-sc-data/byte_code/smartbugs/creation/gpickles/{bug}/clean_{count}_buggy_curated_0/compressed_graphs/cfg_compressed_graphs.gpickle'
# nodes_edges_compressed_graph_correlation(compressed_graph, axs[1, 0], title='CFG creation')
# compressed_graph = f'./experiments/ge-sc-data/byte_code/smartbugs/runtime/gpickles/{bug}/clean_{count}_buggy_curated_0/compressed_graphs/cfg_compressed_graphs.gpickle'
# nodes_edges_compressed_graph_correlation(compressed_graph, axs[1, 1], title='CFG runtime')
# output = f'./forensics/{bug}_nodes_edges_correlation.png'
# fig.tight_layout()
# plt.savefig(output)
# # Move graphs
# from shutil import copy
# bug_type = {'access_control': 57, 'arithmetic': 60, 'denial_of_service': 46,
# 'front_running': 44, 'reentrancy': 71, 'time_manipulation': 50,
# 'unchecked_low_level_calls': 95}
# for bug, count in bug_type.items():
# creation_source = f'./experiments/ge-sc-data/byte_code/smartbugs/creation/gpickles/{bug}/clean_{count}_buggy_curated_0/compressed_graphs/cfg_compressed_graphs.gpickle'
# creation_target = f'./experiments/ge-sc-data/byte_code/smartbugs/creation/gpickles/compressed_graphs/{bug}_cfg_compressed_graphs.gpickle'
# runtime_source = f'./experiments/ge-sc-data/byte_code/smartbugs/runtime/gpickles/{bug}/clean_{count}_buggy_curated_0/compressed_graphs/cfg_compressed_graphs.gpickle'
# runtime_target = f'./experiments/ge-sc-data/byte_code/smartbugs/runtime/gpickles/compressed_graphs/{bug}_cfg_compressed_graphs.gpickle'
# copy(creation_source, creation_target)
# copy(runtime_source, runtime_target)
## Generate historgram
# last_hidden = '/home/minhnn/minhnn/ICSE/ge-sc/forensics/graph_hiddens/reentrancy/last_attention.pt'
# hiddens = torch.load(last_hidden)
# fig, axes = plt.subplots(nrows=2, ncols=2)
# n_bins = np.linspace(torch.min(hiddens).item(), torch.max(hiddens).item(), num=200)
# axes[0, 0].hist(hiddens, n_bins, histtype='step', fill=False)
# axes[0, 0].set_title('stack step (unfilled)')
# fig.tight_layout()
# plt.show()
# ax = plt.figure().add_subplot(projection='3d')
# x = np.linspace(torch.min(hiddens).item(), torch.max(hiddens).item(), num=200)
# lambdas = range(1, hiddens.shape[0]+1)
# # verts[i] is a list of (x, y) pairs defining polygon i.
# verts = [polygon_under_graph(x, poisson.pmf(l, x)) for l in lambdas]
# facecolors = plt.colormaps['viridis_r'](np.linspace(0, 1, len(verts)))
# poly = PolyCollection(verts, facecolors=facecolors, alpha=.7)
# ax.add_collection3d(poly, zs=lambdas, zdir='y')
# ax.set(xlim=(-3, 3), ylim=(0, 40), zlim=(0, 1500),
# xlabel='x', ylabel=r'$\lambda$', zlabel='probability')
# plt.show()
## 3D last hidden visual
# ax = plt.figure().add_subplot(projection='3d')
# x = np.linspace(torch.min(hiddens).item(), torch.max(hiddens).item(), num=200)
# hiddens_hist = [np.histogram(h.numpy(), bins=x, range=None, normed=None, weights=None, density=None)[0].tolist() for h in hiddens]
# # print([np.sum(hid * np.diff(x)) for hid in hiddens_hist])
# print(max(hiddens_hist[0]))
# lambdas = lambdas = range(1, hiddens.shape[0]+1)
# # verts[i] is a list of (x, y) pairs defining polygon i.
# verts = [polygon_under_graph(x, y) for y in hiddens_hist]
# # print(verts)
# # facecolors = plt.colormaps()[0](np.linspace(0, 1, len(verts)))
# viridis = cm.get_cmap('viridis', 8)
# facecolors = viridis(np.linspace(0, 1, len(verts)))
# poly = PolyCollection(verts, facecolors=facecolors, alpha=.6)
# ax.add_collection3d(poly, zs=lambdas, zdir='y')
# ax.set(xlim=(torch.min(hiddens).item(), torch.max(hiddens).item()), ylim=(0, 30), zlim=(0, 7),
# xlabel='x', ylabel='y', zlabel='counter')
# plt.show()
## PCA last hidden state
# plt.figure(figsize=(20, 18), dpi=80)
# models = ['metapath2vec', 'nodetype', 'line', 'node2vec', 'random_32', 'random_64', 'random_128', 'zeros_32', 'zeros_64', 'zeros_128']
# for idx, (bug, count) in enumerate(bug_type.items()):
# fig, axes = plt.subplots(nrows=3, ncols=4, figsize=(20,13), dpi=100)
# for model_idx, model in enumerate(models):
# ax = axes[int(model_idx/4), int(model_idx%4)]
# last_hidden = f'./experiments/logs/graph_classification/byte_code/smartbugs/runtime/han/cfg/{model}/{bug}/last_hiddens.json'
# with open(last_hidden, 'r') as f:
# content = json.load(f)
# hiddens_0 = [ann['hiddens'] for ann in content if ann['targets'] == 0]
# targets_0 = [ann['targets'] for ann in content if ann['targets'] == 0]
# contract_names_0 = [ann['contract_name'] for ann in content if ann['targets'] == 0]
# hiddens_1 = [ann['hiddens'] for ann in content if ann['targets'] == 1]
# targets_1 = [ann['targets'] for ann in content if ann['targets'] == 1]
# contract_names_1 = [ann['contract_name'] for ann in content if ann['targets'] == 1]
# assert len(hiddens_0) + len(hiddens_1) == len(content)
# print(len(hiddens_0)/(len(content)))
# # fig = plt.figure()
# # ax = fig.add_subplot(projection='3d')
# # ax = fig.add_subplot()
# # fig = plt.figure(1, figsize=(8, 6))
# # ax = Axes3D(fig, elev=-150, azim=110)
# # x_3dims = PCA(n_components=3).fit_transform(hiddens)
# hidden_0_2dims = PCA(n_components=2).fit_transform(np.array(hiddens_0))
# hidden_1_2dims = PCA(n_components=2).fit_transform(np.array(hiddens_1))
# # ax.scatter(x_3dims[:, 0], x_3dims[:, 1], x_3dims[:, 2], marker='o')
# ax.scatter(hidden_0_2dims[:, 0], hidden_0_2dims[:, 1], marker='o', label='normal')
# ax.scatter(hidden_1_2dims[:, 0], hidden_1_2dims[:, 1], marker='^', label='buggy')
# ax.set_title(f'{model}', fontsize=15)
# # ax.set_zlabel('Z Label')
# forensic_path = f'./forensics/last_hiddens/runtime/han/{bug}_last_hiddent.png'
# axes[0, 0].legend(ncol=1, bbox_to_anchor=(0, 1), loc='lower center', fontsize=15)
# plt.savefig(forensic_path)
# ax.scatter(
# X_reduced[:, 0],
# X_reduced[:, 1],
# X_reduced[:, 2],
# c=y,
# cmap=plt.cm.Set1,
# edgecolor="k",
# s=40,
# )
# ax.set_title("First three PCA directions")
# ax.set_xlabel("1st eigenvector")
# ax.w_xaxis.set_ticklabels([])
# ax.set_ylabel("2nd eigenvector")
# ax.w_yaxis.set_ticklabels([])
# ax.set_zlabel("3rd eigenvector")
# ax.w_zaxis.set_ticklabels([])
# plt.show()
## Dataset Satatistics
# fig, axes = plt.subplots(nrows=7, ncols=1)
# WIDTH = 0.9
# for idx, (bug, count) in enumerate(bug_type.items()):
# contract_category_path = f'./experiments/ge-sc-data/source_code/{bug}/clean_{count}_buggy_curated_0/source_code_category.json'
# with open(contract_category_path, 'r') as f:
# contract_category = json.load(f)
# # annotation_path = f'./experiments/ge-sc-data/source_code/{bug}/clean_{count}_buggy_curated_0/contract_labels.json'
# annotation_path = f'./experiments/ge-sc-data/byte_code/smartbugs/contract_labels/{bug}/runtime_balanced_contract_labels.json'
# with open(annotation_path, 'r') as f:
# annotations = json.load(f)
# contract_counter = {'curated': {}, 'solidifi': {}, 'clean': {}}
# for record in annotations:
# source_name = record['contract_name'].split('-')[0] + '.sol'
# normal = 1 - record['targets']
# buggy = 1 - normal
# if source_name in list(contract_category['curated'].keys()):
# if source_name not in contract_counter['curated']:
# contract_counter['curated'][source_name] = {'normal': normal, 'buggy': buggy}
# else:
# contract_counter['curated'][source_name]['normal'] += normal
# contract_counter['curated'][source_name]['buggy'] += buggy
# elif source_name in list(contract_category['solidifi'].keys()):
# if source_name not in contract_counter['solidifi']:
# contract_counter['solidifi'][source_name] = {'normal': normal, 'buggy': buggy}
# else:
# contract_counter['solidifi'][source_name]['normal'] += normal
# contract_counter['solidifi'][source_name]['buggy'] += buggy
# else:
# if source_name not in contract_counter['clean']:
# contract_counter['clean'][source_name] = {'normal': normal, 'buggy': buggy}
# else:
# contract_counter['clean'][source_name]['normal'] += normal
# contract_counter['clean'][source_name]['buggy'] += buggy
# total_buggy = 0
# total_contract = 0
# # Curated
# curated_ind = np.arange(len(contract_counter['curated']))
# normal_counter = [contract['normal'] for contract in contract_counter['curated'].values()]
# buggy_counter = [contract['buggy'] for contract in contract_counter['curated'].values()]
# c_normal = axes[idx].bar(curated_ind, normal_counter, WIDTH, label='normal curated')
# c_buggy = axes[idx].bar(curated_ind, buggy_counter, WIDTH, bottom=normal_counter, label='buggy curated')
# curated_text = f'Curated Buggy: {sum(buggy_counter)}/{sum(buggy_counter) + sum(normal_counter)} ({round(sum(buggy_counter)/(sum(buggy_counter) + sum(normal_counter))*100,2)}%)'
# total_buggy += sum(buggy_counter)
# total_contract += sum(buggy_counter) + sum(normal_counter)
# max_y = max(0, max(normal_counter + buggy_counter))
# # Solidifi
# buggy_ind = np.arange(len(curated_ind), len(curated_ind) + len(contract_counter['solidifi']))
# normal_counter = [contract['normal'] for contract in contract_counter['solidifi'].values()]
# buggy_counter = [contract['buggy'] for contract in contract_counter['solidifi'].values()]
# s_normal = axes[idx].bar(buggy_ind, normal_counter, WIDTH, label='normal solidifi')
# s_buggy = axes[idx].bar(buggy_ind, buggy_counter, WIDTH, bottom=normal_counter, label='buggy solidifi')
# solidifi_text = f'Solidifi Buggy: {sum(buggy_counter)}/{sum(buggy_counter) + sum(normal_counter)} ({round(sum(buggy_counter)/(sum(buggy_counter) + sum(normal_counter))*100,2)}%)'
# total_buggy += sum(buggy_counter)
# total_contract += sum(buggy_counter) + sum(normal_counter)
# max_y = max(max_y, max(normal_counter + buggy_counter))
# # Clean
# clean_ind = np.arange(len(curated_ind) + len(buggy_ind), len(curated_ind) + len(buggy_ind) + len(contract_counter['clean']))
# normal_counter = [contract['normal'] for contract in contract_counter['clean'].values()]
# buggy_counter = [contract['buggy'] for contract in contract_counter['clean'].values()]
# cl_normal = axes[idx].bar(clean_ind, normal_counter, WIDTH/3, label='clean')
# total_contract += sum(normal_counter)
# # cl_buggy = axes[idx].bar(clean_ind, buggy_counter, WIDTH, bottom=normal_counter, label='clean')
# total_text = f'Total Buggy: {total_buggy}/{total_contract} ({round(total_buggy/total_contract*100,2)})%)'
# # axes[idx].axhline(0, color='grey', linewidth=0.8)
# axes[idx].set_ylabel(bug[:10])
# max_y = max(max_y, max(normal_counter + buggy_counter + [0]))
# # Label with label_type 'center' instead of the default 'edge'
# axes[idx].bar_label(c_normal, label_type='center', fontsize = 7)
# axes[idx].bar_label(c_buggy, label_type='center', fontsize = 7)
# axes[idx].bar_label(s_normal, label_type='center', fontsize = 7)
# axes[idx].bar_label(s_buggy, label_type='center', fontsize = 7)
# axes[idx].bar_label(cl_normal, label_type='edge', fontsize = 7)
# axes[idx].text(-1, max_y, '\n'.join([curated_text, solidifi_text, total_text]), horizontalalignment='left', verticalalignment='top')
# axes[0].legend(ncol=5, bbox_to_anchor=(0, 1),
# loc='lower left', fontsize=8)
# plt.show()
# Graph statistics
for bug, count in bug_type.items():
compressed_graph = f'./experiments/ge-sc-data/source_code/{bug}/buggy_curated/cfg_cg_compressed_graphs.gpickle'
print(bug, ' : ', count)
nx_graph = nx.read_gpickle(compressed_graph)
print('num of nodes: ', len(nx_graph.nodes()))
print('num of edges: ', len(nx_graph.edges()))
bug_node = 0
for n, data in nx_graph.nodes(data=True):
if data['node_info_vulnerabilities'] is None:
continue
bug_node += 1
print('bug node: ', bug_node)