Secure-MeanShift-Clustering/mod2_distcal.py at main · ASU-CSE598-SC4ML/Secure-MeanShift-Clustering · 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
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
#%%

#Example Usage
# _ , gt_centroid = point_gen([0,1], [0,1], 4, 1000, if_plot = True)
# print("Ground-truth centroid is: ", gt_centroid)

import numpy as np
import torch
from sklearn.cluster import MeanShift
from matplotlib import pyplot as plt
import matplotlib.cm as cm
from mod1_generate import point_gen
import crypten
import crypten.mpc as mpc
import crypten.communicator as comm
import pickle

#Example Usage
    # dist_cal1 = distance_calculation(2, 1, 1)
    # dist_cal1.enc()
    # dist_cal1.discal()

def sample_dust(point_array, n_dusts):
    sample_idx = np.random.choice(point_array.shape[0], n_dusts, replace=False)
    return point_array[sample_idx]

class distance_calculation(object):
    def __init__(self, point_array, n_dust = 1, radius = 0.1, if_plot = True):
        crypten.init()
        torch.set_num_threads(1)
        self.upper_x = 1
        self.lower_x = 0
        self.upper_y = 1
        self.lower_y = 0
        self.n_point = point_array.shape[0]
        self.n_dust = n_dust
        self.radius = radius
        self.point_array = point_array
        # point_gen([self.lower_x,self.upper_x], [self.lower_y,self.upper_y], self.n_center, self.n_point, radius = self.radius, if_plot = if_plot)
        self.dust_array = sample_dust(self.point_array, self.n_dust)
        # if if_plot:
        #     x, y = self.dust_array.T
        #     plt.scatter(x, y, marker="*", s=256, color = "m")
        #     plt.xlim(self.lower_x,  self.upper_x)
        #     plt.ylim(self.lower_y,  self.upper_y)
        #     plt.show()

    @mpc.run_multiprocess(world_size=2) # Two process will run the identical code below:
    def enc(self, verify = False):
        rank = comm.get().get_rank()
        dust_share_list = []
        print(self.dust_array.shape, self.point_array.shape)
        for i in range(self.n_dust):
            dust = list(self.dust_array[i, :])
            dust_enc = crypten.cryptensor(dust, ptype=crypten.ptype.arithmetic)
            dust_share_list.append(dust_enc)
        point_share_list = []
        for i in range(self.n_point):
            point =  list(self.point_array[i, :])
            point_enc = crypten.cryptensor(point, ptype=crypten.ptype.arithmetic)
            point_share_list.append(point_enc)
        if verify:
            if rank == 0:
                print("=========Start of Verification========")
            for i in range(self.n_dust):
                dust_val = dust_share_list[i].get_plain_text()
                if rank == 0:
                    print("Dust to be Encrypted is: ", self.dust_array[i, :])
                    print("Decrypted Dust is: ", dust_val)
            for i in range(self.n_point):
                point_val = point_share_list[i].get_plain_text()
                if rank == 0:
                    print("Point to be Encrypted is: ", self.point_array[i, :])
                    print("Decrypted Point is: ", point_val)
            if rank == 0:
                print("=========End of Verification========")

        # return dust_share_list, point_share_list, save secret share to file
        return_dict1 = {}
        return_dict1["point_share_list_rank{}".format(rank)] = point_share_list
        with open('data_rank_{}.pickle'.format(rank), 'wb') as handle:
            pickle.dump(return_dict1, handle, protocol=pickle.HIGHEST_PROTOCOL)

        return_dict2 = {}
        return_dict2["centroid_share_list_rank{}".format(rank)] = dust_share_list
        with open('centroid_rank_{}.pickle'.format(rank), 'wb') as handle:
            pickle.dump(return_dict2, handle, protocol=pickle.HIGHEST_PROTOCOL)
    @mpc.run_multiprocess(world_size=2)  # Two process will run the identical code below:
    def print_centroid(self):
        rank = comm.get().get_rank()
        with open('centroid_rank_{}.pickle'.format(rank), 'rb') as handle:
            ss_dict = pickle.load(handle)
        centroid_share_list = ss_dict["centroid_share_list_rank{}".format(rank)]
        n_dust = len(centroid_share_list)
        print_out = "centroid: "
        for i in range(n_dust):
            print_out += " {}".format(str(centroid_share_list[i].get_plain_text()))
        if rank == 0:
            print(print_out)

    @mpc.run_multiprocess(world_size=2)  # Two process will run the identical code below:
    def get_plain_centroid(self):
        rank = comm.get().get_rank()
        with open('centroid_rank_{}.pickle'.format(rank), 'rb') as handle:
            ss_dict = pickle.load(handle)
        centroid_share_list = ss_dict["centroid_share_list_rank{}".format(rank)]
        n_dust = len(centroid_share_list)
        plain_centroid = []

        print_out = "centroid: "
        for i in range(n_dust):
            plain_centroid.append(centroid_share_list[i].get_plain_text().tolist())
        if rank == 0:
            with open('plain_centroid.pickle', 'wb') as handle:
                pickle.dump(plain_centroid, handle, protocol=pickle.HIGHEST_PROTOCOL)
    @mpc.run_multiprocess(world_size=2)  # Two process will run the identical code below:
    def discal(self, verify = False):
        rank = comm.get().get_rank()

        #Receive secret share from enc step.
        with open('data_rank_{}.pickle'.format(rank), 'rb') as handle:
            ss_dict = pickle.load(handle)
        point_share_list = ss_dict["point_share_list_rank{}".format(rank)]

        with open('centroid_rank_{}.pickle'.format(rank), 'rb') as handle:
            ss_dict = pickle.load(handle)
        centroid_share_list = ss_dict["centroid_share_list_rank{}".format(rank)]


        n_dust = len(centroid_share_list)
        # Verify the correctness of received shares.
        if verify:
            if rank == 0:
                print("=========Start of Verification========")
            for i in range(n_dust):
                dust_val = centroid_share_list[i].get_plain_text()
                if rank == 0:
                    print("Dust to be Encrypted is: ", self.dust_array[i, :])
                    print("Decrypted Dust is: ", dust_val)
            for i in range(self.n_point):
                point_val = point_share_list[i].get_plain_text()
                if rank == 0:
                    print("Point to be Encrypted is: ", self.point_array[i, :])
                    print("Decrypted Point is: ", point_val)
            if rank == 0:
                print("=========End of Verification========")

        # Verify distance calculation the secret share.
        if verify:
            if rank == 0:
                print("=========Start of Verification========")
            true_distance = sum(((point_share_list[0].get_plain_text() - point_share_list[1].get_plain_text())**2))
            dist_sum = (point_share_list[0] ** 2 +  point_share_list[1] ** 2 - 2 * (point_share_list[0] * point_share_list[1])).sum()
            dist_sum_val = dist_sum.get_plain_text()
            if rank == 0:
                print("Ground-Truth Distance is: ", true_distance)
                print("Decrypted Distance is: ", dist_sum_val)
                print("=========End of Verification========")

        #Calculate Distance
        distance_share_list = []
        for i in range(n_dust):
            tmp_list = []
            for j in range(self.n_point):
                dist_sum = (centroid_share_list[i] ** 2 +  point_share_list[j] ** 2 - 2 * (centroid_share_list[i] * point_share_list[j])).sum()
                tmp_list.append(dist_sum)
            distance_share_list.append(tmp_list)

        #Save Distance
        return_dict = {}
        return_dict["distance_share_list_rank{}".format(rank)] = distance_share_list
        with open('dist_rank_{}.pickle'.format(rank), 'wb') as handle:
            pickle.dump(return_dict, handle, protocol=pickle.HIGHEST_PROTOCOL)

    @mpc.run_multiprocess(world_size=2)  # Two process will run the identical code below:
    def verify_discal(self):
        rank = comm.get().get_rank()

        #Receive secret share from distcal step.
        with open('dist_rank_{}.pickle'.format(rank), 'rb') as handle:
            dist_dict = pickle.load(handle)
        distance_share_list = dist_dict["distance_share_list_rank{}".format(rank)]

        n_dust = len(distance_share_list)
        #Verify each distance
        if rank == 0:
            print("=========Start of Verification========")
        for i in range(n_dust):
            for j in range(self.n_point):
                if rank == 0:
                    gt_dist = sum((self.dust_array[i, :] - self.point_array[j, :]) ** 2)
                    print("Ground-Truth Distance is: ", gt_dist)
                dist_ss = distance_share_list[i][j].get_plain_text()
                if rank == 0:
                    print("Decrypted Distance is: ", dist_ss)
        if rank == 0:
            print("=========End of Verification========")

if __name__ == '__main__':
    pass
    #Example Usage
    # dist_cal1 = distance_calculation()
    # dist_cal1.enc()
    # dist_cal1.discal()
# %%