Implement BinDisc and tests

.gitignore

@@ -31,6 +31,8 @@
 *.out
 *.app
 **/build
+build_Debug
+build_Release
 **/lcoverage
 .idea
 cmake-*

BinDisc.cpp

@@ -0,0 +1,138 @@
+#include <algorithm>
+#include <limits>
+#include <cmath>
+#include "BinDisc.h"
+#include <iostream>
+#include <string>
+
+namespace mdlp {
+
+    BinDisc::BinDisc(int n_bins, strategy_t strategy) : n_bins{ n_bins }, strategy{ strategy }
+    {
+        if (n_bins < 3) {
+            throw std::invalid_argument("n_bins must be greater than 2");
+        }
+    }
+    BinDisc::~BinDisc() = default;
+    void BinDisc::fit(samples_t& X)
+    {
+        cutPoints.clear();
+        if (X.empty()) {
+            cutPoints.push_back(std::numeric_limits<precision_t>::max());
+            return;
+        }
+        if (strategy == strategy_t::QUANTILE) {
+            fit_quantile(X);
+        } else if (strategy == strategy_t::UNIFORM) {
+            fit_uniform(X);
+        }
+    }
+    std::vector<precision_t> linspace(precision_t start, precision_t end, int num)
+    {
+        // Doesn't include end point as it is not needed
+        if (start == end) {
+            return { 0 };
+        }
+        precision_t delta = (end - start) / static_cast<precision_t>(num - 1);
+        std::vector<precision_t> linspc;
+        for (size_t i = 0; i < num - 1; ++i) {
+            precision_t val = start + delta * static_cast<precision_t>(i);
+            linspc.push_back(val);
+        }
+        return linspc;
+    }
+    size_t clip(const size_t n, size_t lower, size_t upper)
+    {
+        return std::max(lower, std::min(n, upper));
+    }
+    std::vector<precision_t> percentile(samples_t& data, std::vector<precision_t>& percentiles)
+    {
+        // Implementation taken from https://dpilger26.github.io/NumCpp/doxygen/html/percentile_8hpp_source.html
+        std::vector<precision_t> results;
+        results.reserve(percentiles.size());
+        for (auto percentile : percentiles) {
+            const size_t i = static_cast<size_t>(std::floor(static_cast<double>(data.size() - 1) * percentile / 100.));
+            const auto indexLower = clip(i, 0, data.size() - 1);
+            const double percentI = static_cast<double>(indexLower) / static_cast<double>(data.size() - 1);
+            const double fraction =
+                (percentile / 100.0 - percentI) /
+                (static_cast<double>(indexLower + 1) / static_cast<double>(data.size() - 1) - percentI);
+            const auto value = data[indexLower] + (data[indexLower + 1] - data[indexLower]) * fraction;
+            if (value != results.back())
+                results.push_back(value);
+        }
+        return results;
+    }
+    void BinDisc::fit_quantile(samples_t& X)
+    {
+        auto quantiles = linspace(0.0, 100.0, n_bins + 1);
+        auto data = X;
+        std::sort(data.begin(), data.end());
+        if (data.front() == data.back() || data.size() == 1) {
+            // if X is constant
+            cutPoints.push_back(std::numeric_limits<precision_t>::max());
+            return;
+        }
+        cutPoints = percentile(data, quantiles);
+        normalizeCutPoints();
+    }
+    void BinDisc::fit_uniform(samples_t& X)
+    {
+
+        auto minmax = std::minmax_element(X.begin(), X.end());
+        cutPoints = linspace(*minmax.first, *minmax.second, n_bins + 1);
+        normalizeCutPoints();
+    }
+    void BinDisc::normalizeCutPoints()
+    {
+        // Add max value to the end
+        cutPoints.push_back(std::numeric_limits<precision_t>::max());
+        // Remove first as it is not needed
+        cutPoints.erase(cutPoints.begin());
+    }
+    labels_t& BinDisc::transform(const samples_t& X)
+    {
+        discretizedData.clear();
+        discretizedData.reserve(X.size());
+        for (const precision_t& item : X) {
+            auto upper = std::upper_bound(cutPoints.begin(), cutPoints.end(), item);
+            discretizedData.push_back(upper - cutPoints.begin());
+        }
+        return discretizedData;
+    }
+}
+// void BinDisc::fit_quantile(samples_t& X)
+    // {
+    //     cutPoints.clear();
+    //     if (X.empty()) {
+    //         cutPoints.push_back(std::numeric_limits<float>::max());
+    //         return;
+    //     }
+    //     samples_t data = X;
+    //     std::sort(data.begin(), data.end());
+    //     float min_val = data.front();
+    //     float max_val = data.back();
+    //     // Handle case of all data points having the same value
+    //     if (min_val == max_val) {
+    //         cutPoints.push_back(std::numeric_limits<float>::max());
+    //         return;
+    //     }
+    //     int first = X.size() / n_bins;
+    //     cutPoints.push_back(data.at(first - 1));
+    //     int bins_done = 1;
+    //     int prev = first - 1;
+    //     while (bins_done < n_bins) {
+    //         int next = first * (bins_done + 1) - 1;
+    //         while (next < X.size() && data.at(next) == data[prev]) {
+    //             ++next;
+    //         }
+    //         if (next == X.size() || bins_done == n_bins - 1) {
+    //             cutPoints.push_back(std::numeric_limits<float>::max());
+    //             break;
+    //         } else {
+    //             cutPoints.push_back(data[next]);
+    //             bins_done++;
+    //             prev = next;
+    //         }
+    //     }
+    // }

BinDisc.h

@@ -0,0 +1,31 @@
+#ifndef BINDISC_H
+#define BINDISC_H
+
+#include "typesFImdlp.h"
+#include <string>
+
+namespace mdlp {
+
+    enum class strategy_t {
+        UNIFORM,
+        QUANTILE
+    };
+    class BinDisc {
+    public:
+        BinDisc(int n_bins = 3, strategy_t strategy = strategy_t::UNIFORM);
+        ~BinDisc();
+        void fit(samples_t&);
+        inline cutPoints_t getCutPoints() const { return cutPoints; };
+        labels_t& transform(const samples_t&);
+        static inline std::string version() { return "1.0.0"; };
+    private:
+        void fit_uniform(samples_t&);
+        void fit_quantile(samples_t&);
+        void normalizeCutPoints();
+        int n_bins;
+        strategy_t strategy;
+        labels_t discretizedData = labels_t();
+        cutPoints_t cutPoints;
+    };
+}
+#endif

CPPFImdlp.cpp

@@ -3,7 +3,6 @@
 #include <set>
 #include <cmath>
 #include "CPPFImdlp.h"
-#include "Metrics.h"
 
 namespace mdlp {
 
@@ -178,7 +177,7 @@ namespace mdlp {
     indices_t CPPFImdlp::sortIndices(samples_t& X_, labels_t& y_)
     {
         indices_t idx(X_.size());
-        iota(idx.begin(), idx.end(), 0);
+        std::iota(idx.begin(), idx.end(), 0);
         stable_sort(idx.begin(), idx.end(), [&X_, &y_](size_t i1, size_t i2) {
             if (X_[i1] == X_[i2])
                 return y_[i1] < y_[i2];
@@ -214,7 +213,7 @@ namespace mdlp {
         discretizedData.clear();
         discretizedData.reserve(data.size());
         for (const precision_t& item : data) {
-            auto upper = upper_bound(cutPoints.begin(), cutPoints.end(), item);
+            auto upper = std::upper_bound(cutPoints.begin(), cutPoints.end(), item);
             discretizedData.push_back(upper - cutPoints.begin());
         }
         return discretizedData;

CPPFImdlp.h

@@ -2,13 +2,22 @@
 #define CPPFIMDLP_H
 
 #include "typesFImdlp.h"
-#include "Metrics.h"
 #include <limits>
 #include <utility>
 #include <string>
+#include "Metrics.h"
 
 namespace mdlp {
     class CPPFImdlp {
+    public:
+        CPPFImdlp();
+        CPPFImdlp(size_t, int, float);
+        ~CPPFImdlp();
+        void fit(samples_t&, labels_t&);
+        inline cutPoints_t getCutPoints() const { return cutPoints; };
+        labels_t& transform(const samples_t&);
+        inline int get_depth() const { return depth; };
+        static inline std::string version() { return "1.1.3"; };
     protected:
         size_t min_length = 3;
         int depth = 0;
@@ -21,25 +30,13 @@ namespace mdlp {
         cutPoints_t cutPoints;
         size_t num_cut_points = numeric_limits<size_t>::max();
         labels_t discretizedData = labels_t();
-
         static indices_t sortIndices(samples_t&, labels_t&);
-
         void computeCutPoints(size_t, size_t, int);
         void resizeCutPoints();
         bool mdlp(size_t, size_t, size_t);
         size_t getCandidate(size_t, size_t);
         size_t compute_max_num_cut_points() const;
         pair<precision_t, size_t> valueCutPoint(size_t, size_t, size_t);
-
-    public:
-        CPPFImdlp();
-        CPPFImdlp(size_t, int, float);
-        ~CPPFImdlp();
-        void fit(samples_t&, labels_t&);
-        inline cutPoints_t getCutPoints() const { return cutPoints; };
-        labels_t& transform(const samples_t&);
-        inline int get_depth() const { return depth; };
-        static inline string version() { return "1.1.2"; };
     };
 }
 #endif

sonar-project.properties

@@ -3,7 +3,7 @@ sonar.organization=rmontanana
 
 # This is the name and version displayed in the SonarCloud UI.
 sonar.projectName=mdlp
-sonar.projectVersion=1.0.2
+sonar.projectVersion=1.1.3
 # sonar.test.exclusions=tests/**
 # sonar.tests=tests/
 # sonar.coverage.exclusions=tests/**,sample/**

tests/BinDisc_unittest.cpp

@@ -0,0 +1,351 @@
+#include <fstream>
+#include <string>
+#include <iostream>
+#include "gtest/gtest.h"
+#include "ArffFiles.h"
+#include "../BinDisc.h"
+
+namespace mdlp {
+    const float margin = 1e-4;
+    static std::string set_data_path()
+    {
+        std::string path = "../datasets/";
+        std::ifstream file(path + "iris.arff");
+        if (file.is_open()) {
+            file.close();
+            return path;
+        }
+        return "../../tests/datasets/";
+    }
+    const std::string data_path = set_data_path();
+    class TestBinDisc3U : public BinDisc, public testing::Test {
+    public:
+        TestBinDisc3U(int n_bins = 3) : BinDisc(n_bins, strategy_t::UNIFORM) {};
+    };
+    class TestBinDisc3Q : public BinDisc, public testing::Test {
+    public:
+        TestBinDisc3Q(int n_bins = 3) : BinDisc(n_bins, strategy_t::QUANTILE) {};
+    };
+    class TestBinDisc4U : public BinDisc, public testing::Test {
+    public:
+        TestBinDisc4U(int n_bins = 4) : BinDisc(n_bins, strategy_t::UNIFORM) {};
+    };
+    class TestBinDisc4Q : public BinDisc, public testing::Test {
+    public:
+        TestBinDisc4Q(int n_bins = 4) : BinDisc(n_bins, strategy_t::QUANTILE) {};
+    };
+    TEST_F(TestBinDisc3U, Easy3BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_NEAR(3.66667, cuts[0], margin);
+        EXPECT_NEAR(6.33333, cuts[1], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3Q, Easy3BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_NEAR(3.666667, cuts[0], margin);
+        EXPECT_NEAR(6.333333, cuts[1], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3U, X10BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.0, cuts[0]);
+        EXPECT_EQ(7.0, cuts[1]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3Q, X10BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4, cuts[0]);
+        EXPECT_EQ(7, cuts[1]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3U, X11BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_NEAR(4.33333, cuts[0], margin);
+        EXPECT_NEAR(7.66667, cuts[1], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3U, X11BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_NEAR(4.33333, cuts[0], margin);
+        EXPECT_NEAR(7.66667, cuts[1], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3U, ConstantUniform)
+    {
+        samples_t X = { 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[0]);
+        EXPECT_EQ(1, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 0, 0 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3Q, ConstantQuantile)
+    {
+        samples_t X = { 1.0, 1.0, 1.0, 1.0, 1.0, 1.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[0]);
+        EXPECT_EQ(1, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 0, 0 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc3U, EmptyUniform)
+    {
+        samples_t X = {};
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[0]);
+        EXPECT_EQ(1, cuts.size());
+    }
+    TEST_F(TestBinDisc3Q, EmptyQuantile)
+    {
+        samples_t X = {};
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[0]);
+        EXPECT_EQ(1, cuts.size());
+    }
+    TEST(TestBinDisc3, ExceptionNumberBins)
+    {
+        EXPECT_THROW(BinDisc(2), std::invalid_argument);
+    }
+    TEST_F(TestBinDisc3U, EasyRepeated)
+    {
+        samples_t X = { 3.0, 1.0, 1.0, 3.0, 1.0, 1.0, 3.0, 1.0, 1.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_NEAR(1.66667, cuts[0], margin);
+        EXPECT_NEAR(2.33333, cuts[1], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 2, 0, 0, 2, 0, 0, 2, 0, 0 };
+        EXPECT_EQ(expected, labels);
+        EXPECT_EQ(3.0, X[0]); // X is not modified
+    }
+    TEST_F(TestBinDisc3Q, EasyRepeated)
+    {
+        samples_t X = { 3.0, 1.0, 1.0, 3.0, 1.0, 1.0, 3.0, 1.0, 1.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        std::cout << "cuts: ";
+        for (auto cut : cuts) {
+            std::cout << cut << " ";
+        }
+        std::cout << std::endl;
+        std::cout << std::string(80, '-') << std::endl;
+        EXPECT_NEAR(1.66667, cuts[0], margin);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[1]);
+        EXPECT_EQ(2, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 1, 0, 0, 1, 0, 0, 1, 0, 0 };
+        EXPECT_EQ(expected, labels);
+        EXPECT_EQ(3.0, X[0]); // X is not modified
+    }
+    TEST_F(TestBinDisc4U, Easy4BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(3.75, cuts[0]);
+        EXPECT_EQ(6.5, cuts[1]);
+        EXPECT_EQ(9.25, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4Q, Easy4BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(3.75, cuts[0]);
+        EXPECT_EQ(6.5, cuts[1]);
+        EXPECT_EQ(9.25, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4U, X13BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.0, cuts[0]);
+        EXPECT_EQ(7.0, cuts[1]);
+        EXPECT_EQ(10.0, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4Q, X13BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.0, cuts[0]);
+        EXPECT_EQ(7.0, cuts[1]);
+        EXPECT_EQ(10.0, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4U, X14BinsUniform)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.25, cuts[0]);
+        EXPECT_EQ(7.5, cuts[1]);
+        EXPECT_EQ(10.75, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4Q, X14BinsQuantile)
+    {
+        samples_t X = { 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.25, cuts[0]);
+        EXPECT_EQ(7.5, cuts[1]);
+        EXPECT_EQ(10.75, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4U, X15BinsUniform)
+    {
+        samples_t X = { 15.0, 8.0, 12.0, 14.0, 6.0, 1.0, 13.0, 11.0, 10.0, 9.0, 7.0, 4.0, 3.0, 5.0, 2.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.5, cuts[0]);
+        EXPECT_EQ(8, cuts[1]);
+        EXPECT_EQ(11.5, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 3, 2, 3, 3, 1, 0, 3, 2, 2, 2, 1, 0, 0, 1, 0 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4Q, X15BinsQuantile)
+    {
+        samples_t X = { 15.0, 13.0, 12.0, 14.0, 6.0, 1.0, 8.0, 11.0, 10.0, 9.0, 7.0, 4.0, 3.0, 5.0, 2.0 };
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(4.5, cuts[0]);
+        EXPECT_EQ(8, cuts[1]);
+        EXPECT_EQ(11.5, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 3, 3, 3, 3, 1, 0, 2, 2, 2, 2, 1, 0, 0, 1, 0 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4U, RepeatedValuesUniform)
+    {
+        samples_t X = { 0.0, 1.0, 1.0, 1.0, 2.0, 2.0, 3.0, 3.0, 3.0, 4.0 };
+        //               0    1     2   3    4    5    6    7    8    9
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(1.0, cuts[0]);
+        EXPECT_EQ(2.0, cuts[1]);
+        EXPECT_EQ(3.0, cuts[2]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[3]);
+        EXPECT_EQ(4, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 1, 1, 1, 2, 2, 3, 3, 3, 3 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4Q, RepeatedValuesQuantile)
+    {
+        samples_t X = { 0.0, 1.0, 1.0, 1.0, 2.0, 2.0, 3.0, 3.0, 3.0, 4.0 };
+        //               0    1     2   3    4    5    6    7    8    9
+        fit(X);
+        auto cuts = getCutPoints();
+        EXPECT_EQ(2.0, cuts[0]);
+        EXPECT_EQ(3.0, cuts[1]);
+        EXPECT_EQ(numeric_limits<float>::max(), cuts[2]);
+        EXPECT_EQ(3, cuts.size());
+        auto labels = transform(X);
+        labels_t expected = { 0, 0, 0, 0, 1, 1, 2, 2, 2, 2 };
+        EXPECT_EQ(expected, labels);
+    }
+    TEST_F(TestBinDisc4U, irisUniform)
+    {
+        ArffFiles file;
+        file.load(data_path + "iris.arff", true);
+        vector<samples_t>& X = file.getX();
+        fit(X[0]);
+        auto Xt = transform(X[0]);
+        labels_t expected = { 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 3, 2, 2, 1, 2, 1, 2, 0, 2, 0, 0, 1, 1, 1, 1, 2, 1, 1, 2, 1, 1, 1, 2, 1, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 1, 1, 1, 1, 1, 0, 1, 1, 1, 2, 0, 1, 2, 1, 3, 2, 2, 3, 0, 3, 2, 3, 2, 2, 2, 1, 1, 2, 2, 3, 3, 1, 2, 1, 3, 2, 2, 3, 2, 1, 2, 3, 3, 3, 2, 2, 1, 3, 2, 2, 1, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2, 1 };
+        EXPECT_EQ(expected, Xt);
+    }
+    TEST_F(TestBinDisc4Q, irisQuantile)
+    {
+        ArffFiles file;
+        file.load(data_path + "iris.arff", true);
+        vector<samples_t>& X = file.getX();
+        fit(X[0]);
+        auto Xt = transform(X[0]);
+        labels_t expected = { 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 2, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 3, 3, 3, 1, 3, 1, 2, 0, 3, 1, 0, 2, 2, 2, 1, 3, 1, 2, 2, 1, 2, 2, 2, 2, 3, 3, 3, 3, 2, 1, 1, 1, 2, 2, 1, 2, 3, 2, 1, 1, 1, 2, 2, 0, 1, 1, 1, 2, 1, 1, 2, 2, 3, 2, 3, 3, 0, 3, 3, 3, 3, 3, 3, 1, 2, 3, 3, 3, 3, 2, 3, 1, 3, 2, 3, 3, 2, 2, 3, 3, 3, 3, 3, 2, 2, 3, 2, 3, 2, 3, 3, 3, 2, 3, 3, 3, 2, 3, 2, 2 };
+        EXPECT_EQ(expected, Xt);
+    }
+}

tests/CMakeLists.txt

@@ -1,3 +1,4 @@
+cmake_minimum_required(VERSION 3.20)
 set(CMAKE_CXX_STANDARD 11)
 include(FetchContent)
 
@@ -16,14 +17,18 @@ enable_testing()
 
 add_executable(Metrics_unittest ../Metrics.cpp Metrics_unittest.cpp)
 add_executable(FImdlp_unittest ../CPPFImdlp.cpp ArffFiles.cpp ../Metrics.cpp FImdlp_unittest.cpp)
+add_executable(BinDisc_unittest ../BinDisc.cpp ArffFiles.cpp BinDisc_unittest.cpp)
 target_link_libraries(Metrics_unittest GTest::gtest_main)
 target_link_libraries(FImdlp_unittest GTest::gtest_main)
+target_link_libraries(BinDisc_unittest GTest::gtest_main)
 target_compile_options(Metrics_unittest PRIVATE --coverage)
 target_compile_options(FImdlp_unittest PRIVATE --coverage)
+target_compile_options(BinDisc_unittest PRIVATE --coverage)
 target_link_options(Metrics_unittest PRIVATE --coverage)
 target_link_options(FImdlp_unittest PRIVATE --coverage)
+target_link_options(BinDisc_unittest PRIVATE --coverage)
 
 include(GoogleTest)
 gtest_discover_tests(Metrics_unittest)
 gtest_discover_tests(FImdlp_unittest)
-
+gtest_discover_tests(BinDisc_unittest)

tests/test

@@ -1,3 +1,4 @@
+#!/bin/bash
 if [ -d build ] ; then
 	rm -fr build
 fi
@@ -9,12 +10,9 @@ cmake --build build
 cd build
 ctest --output-on-failure
 cd ..
-if [ ! -d gcovr-report ] ; then 
+mkdir gcovr-report
-	mkdir gcovr-report
-fi
-rm -fr gcovr-report/* 2>/dev/null
 #lcov --capture --directory ./ --output-file lcoverage/main_coverage.info
 #lcov --remove lcoverage/main_coverage.info 'v1/*' '/Applications/*' '*/tests/*' --output-file lcoverage/main_coverage.info -q
 #lcov --list lcoverage/main_coverage.info
 cd ..
-gcovr  --gcov-filter "CPPFImdlp.cpp" --gcov-filter "Metrics.cpp" --txt --sonarqube=tests/gcovr-report/coverage.xml   
+gcovr  --gcov-filter "CPPFImdlp.cpp" --gcov-filter "Metrics.cpp" --gcov-filter "BinDisc.cpp" --txt --sonarqube=tests/gcovr-report/coverage.xml --exclude-noncode-lines

tests/testKbins.py

@@ -0,0 +1,404 @@
+from scipy.io.arff import loadarff
+from sklearn.preprocessing import KBinsDiscretizer
+
+
+def test(clf, X, expected, title):
+    X = [[x] for x in X]
+    clf.fit(X)
+    computed = [int(x[0]) for x in clf.transform(X)]
+    print(f"{title}")
+    print(f"{computed=}")
+    print(f"{expected=}")
+    assert computed == expected
+    print("-" * 80)
+
+
+# Test Uniform Strategy
+clf3u = KBinsDiscretizer(
+    n_bins=3, encode="ordinal", strategy="uniform", subsample=200_000
+)
+clf3q = KBinsDiscretizer(
+    n_bins=3, encode="ordinal", strategy="quantile", subsample=200_000
+)
+clf4u = KBinsDiscretizer(
+    n_bins=4, encode="ordinal", strategy="uniform", subsample=200_000
+)
+clf4q = KBinsDiscretizer(
+    n_bins=4, encode="ordinal", strategy="quantile", subsample=200_000
+)
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0]
+labels = [0, 0, 0, 1, 1, 1, 2, 2, 2]
+test(clf3u, X, labels, title="Easy3BinsUniform")
+test(clf3q, X, labels, title="Easy3BinsQuantile")
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0]
+labels = [0, 0, 0, 1, 1, 1, 2, 2, 2, 2]
+# En C++ se obtiene el mismo resultado en ambos, no como aquí
+labels2 = [0, 0, 0, 1, 1, 1, 1, 2, 2, 2]
+test(clf3u, X, labels, title="X10BinsUniform")
+test(clf3q, X, labels2, title="X10BinsQuantile")
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0]
+labels = [0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 2]
+# En C++ se obtiene el mismo resultado en ambos, no como aquí
+# labels2 = [0, 0, 0, 1, 1, 1, 1, 2, 2, 2]
+test(clf3u, X, labels, title="X11BinsUniform")
+test(clf3q, X, labels, title="X11BinsQuantile")
+#
+X = [1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
+labels = [0, 0, 0, 0, 0, 0]
+test(clf3u, X, labels, title="ConstantUniform")
+test(clf3q, X, labels, title="ConstantQuantile")
+#
+X = [3.0, 1.0, 1.0, 3.0, 1.0, 1.0, 3.0, 1.0, 1.0]
+labels = [2, 0, 0, 2, 0, 0, 2, 0, 0]
+labels2 = [1, 0, 0, 1, 0, 0, 1, 0, 0]  # igual que en C++
+test(clf3u, X, labels, title="EasyRepeatedUniform")
+test(clf3q, X, labels2, title="EasyRepeatedQuantile")
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0]
+labels = [0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3]
+test(clf4u, X, labels, title="Easy4BinsUniform")
+test(clf4q, X, labels, title="Easy4BinsQuantile")
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0]
+labels = [0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3]
+test(clf4u, X, labels, title="X13BinsUniform")
+test(clf4q, X, labels, title="X13BinsQuantile")
+#
+X = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0]
+labels = [0, 0, 0, 0, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3]
+test(clf4u, X, labels, title="X14BinsUniform")
+test(clf4q, X, labels, title="X14BinsQuantile")
+#
+X1 = [15.0, 8.0, 12.0, 14.0, 6.0, 1.0, 13.0, 11.0, 10.0, 9.0, 7.0, 4.0, 3.0, 5.0, 2.0]
+X2 = [15.0, 13.0, 12.0, 14.0, 6.0, 1.0, 8.0, 11.0, 10.0, 9.0, 7.0, 4.0, 3.0, 5.0, 2.0]
+labels1 = [3, 2, 3, 3, 1, 0, 3, 2, 2, 2, 1, 0, 0, 1, 0]
+labels2 = [3, 3, 3, 3, 1, 0, 2, 2, 2, 2, 1, 0, 0, 1, 0]
+test(clf4u, X1, labels1, title="X15BinsUniform")
+test(clf4q, X2, labels2, title="X15BinsQuantile")
+#
+X = [0.0, 1.0, 1.0, 1.0, 2.0, 2.0, 3.0, 3.0, 3.0, 4.0]
+labels = [0, 1, 1, 1, 2, 2, 3, 3, 3, 3]
+test(clf4u, X, labels, title="RepeatedValuesUniform")
+test(clf4q, X, labels, title="RepeatedValuesQuantile")
+
+print(f"Uniform   {clf4u.bin_edges_=}")
+print(f"Quaintile {clf4q.bin_edges_=}")
+print("-" * 80)
+#
+data, meta = loadarff("tests/datasets/iris.arff")
+labelsu = [
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    1,
+    1,
+    1,
+    0,
+    1,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    1,
+    0,
+    0,
+    1,
+    1,
+    1,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    3,
+    2,
+    2,
+    1,
+    2,
+    1,
+    2,
+    0,
+    2,
+    1,
+    0,
+    1,
+    1,
+    2,
+    1,
+    2,
+    1,
+    1,
+    2,
+    1,
+    1,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    2,
+    2,
+    1,
+    1,
+    1,
+    2,
+    1,
+    0,
+    1,
+    1,
+    1,
+    2,
+    0,
+    1,
+    2,
+    1,
+    3,
+    2,
+    2,
+    3,
+    0,
+    3,
+    2,
+    3,
+    2,
+    2,
+    2,
+    1,
+    1,
+    2,
+    2,
+    3,
+    3,
+    1,
+    2,
+    1,
+    3,
+    2,
+    2,
+    3,
+    2,
+    2,
+    2,
+    3,
+    3,
+    3,
+    2,
+    2,
+    2,
+    3,
+    2,
+    2,
+    1,
+    2,
+    2,
+    2,
+    1,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2,
+    1,
+]
+labelsq = [
+    1,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    2,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    1,
+    0,
+    1,
+    0,
+    0,
+    0,
+    1,
+    1,
+    0,
+    0,
+    1,
+    1,
+    1,
+    0,
+    0,
+    1,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    1,
+    0,
+    1,
+    0,
+    3,
+    3,
+    3,
+    1,
+    3,
+    1,
+    2,
+    0,
+    3,
+    1,
+    0,
+    2,
+    2,
+    2,
+    1,
+    3,
+    1,
+    2,
+    2,
+    1,
+    2,
+    2,
+    2,
+    2,
+    3,
+    3,
+    3,
+    3,
+    2,
+    1,
+    1,
+    1,
+    2,
+    2,
+    1,
+    2,
+    3,
+    2,
+    1,
+    1,
+    1,
+    2,
+    2,
+    0,
+    1,
+    1,
+    1,
+    2,
+    1,
+    1,
+    2,
+    2,
+    3,
+    2,
+    3,
+    3,
+    0,
+    3,
+    3,
+    3,
+    3,
+    3,
+    3,
+    1,
+    2,
+    3,
+    3,
+    3,
+    3,
+    2,
+    3,
+    1,
+    3,
+    2,
+    3,
+    3,
+    2,
+    2,
+    3,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2,
+    3,
+    2,
+    3,
+    2,
+    3,
+    3,
+    3,
+    2,
+    3,
+    3,
+    3,
+    2,
+    3,
+    2,
+    2,
+]
+test(clf4u, data["sepallength"], labelsu, title="IrisUniform")
+test(clf4q, data["sepallength"], labelsq, title="IrisQuantile")
+# print("Labels")
+# print(labels)
+# print("Expected")
+# print(expected)
+# for i in range(len(labels)):
+#     if labels[i] != expected[i]:
+#         print(f"Error at {i} {labels[i]} != {expected[i]}")

typesFImdlp.h

@@ -8,11 +8,11 @@
 using namespace std;
 namespace mdlp {
     typedef float precision_t;
-    typedef vector<precision_t> samples_t;
+    typedef std::vector<precision_t> samples_t;
-    typedef vector<int> labels_t;
+    typedef std::vector<int> labels_t;
-    typedef vector<size_t> indices_t;
+    typedef std::vector<size_t> indices_t;
-    typedef vector<precision_t> cutPoints_t;
+    typedef std::vector<precision_t> cutPoints_t;
-    typedef map<pair<int, int>, precision_t> cacheEnt_t;
+    typedef std::map<std::pair<int, int>, precision_t> cacheEnt_t;
-    typedef map<tuple<int, int, int>, precision_t> cacheIg_t;
+    typedef std::map<std::tuple<int, int, int>, precision_t> cacheIg_t;
 }
 #endif