Hugging Face's logo

Spaces:
w4nn4b3M4ST3R
/
DeduplicationImage
Sleeping

App Files Community
DeduplicationImage
File size: 4,746 Bytes 
daea7f9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
#include "min_hash.h"

MinHash::MinHash(size_t numPlanes, size_t dimension)
    : numPlanes(numPlanes),
      dimension(dimension),
      mean(dimension, 0.0),
      median(dimension, 0.0),
      stddev(dimension, 1.0)
{
    randomHyperplanes.resize(numPlanes, std::vector<double>(dimension));
    generateRandomHyperplanes();
}

void MinHash::generateRandomHyperplanes()

{
    std::random_device rd;
    std::mt19937 gen(rd());
    std::normal_distribution<double> dist(0.0, 1.0);

    for (size_t i = 0; i < numPlanes; ++i)
    {
        double norm = 0.0;
        for (size_t j = 0; j < dimension; ++j)
        {
            randomHyperplanes[i][j] = dist(gen);
            norm += randomHyperplanes[i][j] * randomHyperplanes[i][j];
        }
        norm = std::sqrt(norm);
        if (norm == 0.0)
            norm = 1e-12;
        for (size_t j = 0; j < dimension; ++j)
            randomHyperplanes[i][j] /= norm;
    }
}

std::vector<std::vector<uint8_t>>
MinHash::computeSignatures(const std::vector<std::vector<double>> &matrix,
                           bool useMedianThreshold)
{
    size_t n = matrix.size();
    if (n == 0)
        return {};

    std::vector<std::vector<double>> v(n, std::vector<double>(dimension));

    for (size_t i = 0; i < n; ++i)
    {
        const auto &row = matrix[i];
        for (size_t j = 0; j < dimension; ++j)
        {
            double minval = std::numeric_limits<double>::infinity();
            for (size_t p = 0; p < numPlanes; ++p)
            {
                double prod = row[j] * randomHyperplanes[p][j];
                if (prod < minval)
                    minval = prod;
            }
            v[i][j] = minval;
        }
    }

    std::vector<double> tmp;
    tmp.reserve(n);

    for (size_t j = 0; j < dimension; ++j)
    {
        tmp.clear();
        for (size_t i = 0; i < n; ++i)
            tmp.push_back(v[i][j]);

        // mean
        double s = 0.0;
        for (double x : tmp)
            s += x;
        mean[j] = s / static_cast<double>(n);

        // median
        std::nth_element(tmp.begin(), tmp.begin() + tmp.size() / 2, tmp.end());
        double med = tmp[tmp.size() / 2];
        if (tmp.size() % 2 == 0)
        {
            double other = *std::max_element(tmp.begin(), tmp.begin() + tmp.size() / 2);
            med = 0.5 * (med + other);
        }
        median[j] = med;

        // stddev
        double var = 0.0;
        for (double x : tmp)
        {
            double d = x - mean[j];
            var += d * d;
        }
        var /= static_cast<double>(n);
        stddev[j] = std::sqrt(var);
        if (stddev[j] < 1e-12)
            stddev[j] = 1e-12;
    }

    std::vector<std::vector<uint8_t>> signatures(n, std::vector<uint8_t>(dimension, 0));

    for (size_t i = 0; i < n; ++i)
    {
        for (size_t j = 0; j < dimension; ++j)
        {
            bool bit;
            if (useMedianThreshold)
                bit = (v[i][j] >= median[j]);
            else
            {
                double z = (v[i][j] - mean[j]) / stddev[j];
                bit = (z > 0.0);
            }
            signatures[i][j] = static_cast<uint8_t>(bit);
        }
    }

    return signatures;
}

std::vector<uint8_t>

MinHash::hashFunction(const std::vector<double> &vec,

                      bool useMedianThreshold) const

{
    if (median.empty() || mean.empty() || stddev.empty())
        throw std::runtime_error("Must call computeSignatures() before hashFunction()");

    std::vector<double> minVals(dimension);
    for (size_t j = 0; j < dimension; ++j)
    {
        double minval = std::numeric_limits<double>::infinity();
        for (size_t p = 0; p < numPlanes; ++p)
        {
            double prod = vec[j] * randomHyperplanes[p][j];
            if (prod < minval)
                minval = prod;
        }
        minVals[j] = minval;
    }

    std::vector<uint8_t> bits(dimension, 0);
    for (size_t j = 0; j < dimension; ++j)
    {
        bool bit;
        if (useMedianThreshold)
            bit = (minVals[j] >= median[j]);
        else
        {
            double z = (minVals[j] - mean[j]) / stddev[j];
            bit = (z > 0.0);
        }
        bits[j] = static_cast<uint8_t>(bit);
    }

    return bits;
}

// Getters
const std::vector<std::vector<double>> &MinHash::getHyperplanes() const
{
    return randomHyperplanes;
}
const std::vector<double> &MinHash::getMean() const { return mean; }
const std::vector<double> &MinHash::getMedian() const { return median; }
const std::vector<double> &MinHash::getStddev() const { return stddev; }