Add debug output and precision

fimdlp/CPPFImdlp.cpp

@@ -3,35 +3,45 @@
 #include <iostream>
 namespace CPPFImdlp
 {
-    CPPFImdlp::CPPFImdlp()
+    CPPFImdlp::CPPFImdlp() : debug(false), precision(6)
     {
+        divider = pow(10, precision);
+    }
+    CPPFImdlp::CPPFImdlp(int precision, bool debug) : debug(debug), precision(precision)
+    {
+        divider = pow(10, precision);
     }
     CPPFImdlp::~CPPFImdlp()
     {
     }
-    std::vector<double> CPPFImdlp::cutPoints(std::vector<float> &X, std::vector<int> &y)
+    std::vector<float> CPPFImdlp::cutPoints(std::vector<float> &X, std::vector<int> &y)
     {
-        std::vector<double> cutPts;
-        double antx;
-        // int anty;
+        std::vector<float> cutPts;
+        float antx, cutPoint;
+        int anty;
         std::vector<size_t> indices = sortIndices(X);
         antx = X.at(indices[0]);
-        // anty = y.at(indices[0]);
+        anty = y.at(indices[0]);
         for (auto index = indices.begin(); index != indices.end(); ++index)
         {
             // std::cout << X.at(*index) << " -> " << y.at(*index) << " // ";
             //  Definition 2 Cut points are always on boundaries
-            // if (y.at(*index) != anty && antx < X.at(*index))
+            if (y.at(*index) != anty && antx < X.at(*index))
             //  Weka implementation
-            if (antx < X.at(*index))
+            // if (antx < X.at(*index))
             {
-                // std::cout << "* (" << X.at(*index) << ", " << antx << ") // ";
-                cutPts.push_back((X.at(*index) + antx) / 2);
-                // anty = y.at(*index);
+                cutPoint = round((X.at(*index) + antx) / 2 * divider) / divider;
+                if (debug)
+                {
+                    std::cout << "Cut point: " << (antx + X.at(*index)) / 2 << " //";
+                    std::cout << X.at(*index) << " -> " << y.at(*index) << " anty= " << anty;
+                    std::cout << "* (" << X.at(*index) << ", " << antx << ")=" << ((X.at(*index) + antx) / 2) << std::endl;
+                }
+                cutPts.push_back(cutPoint);
             }
             antx = X.at(*index);
+            anty = y.at(*index);
         }
-        // std::cout << std::endl;
         return cutPts;
     }
     std::vector<size_t> CPPFImdlp::sortIndices(std::vector<float> &X)

fimdlp/CPPFImdlp.h

@@ -8,12 +8,17 @@ namespace CPPFImdlp
     class CPPFImdlp
     {
     private:
-        std::vector<size_t> sortIndices(std::vector<float> &);
+        bool debug;
+        int precision;
+        float divider;
+        std::vector<size_t>
+        sortIndices(std::vector<float> &);
 
     public:
         CPPFImdlp();
+        CPPFImdlp(int, bool debug = false);
         ~CPPFImdlp();
-        std::vector<double> cutPoints(std::vector<float> &, std::vector<int> &);
+        std::vector<float> cutPoints(std::vector<float> &, std::vector<int> &);
     };
 }
 #endif

fimdlp/cfimdlp.pyx

@@ -1,16 +1,18 @@
 # distutils: language = c++
 # cython: language_level = 3
 from libcpp.vector cimport vector
+from libcpp cimport bool
 
 cdef extern from "CPPFImdlp.h" namespace "CPPFImdlp":
     cdef cppclass CPPFImdlp:
         CPPFImdlp() except + 
-        vector[double] cutPoints(vector[float]&, vector[int]&)
+        CPPFImdlp(int, bool) except + 
+        vector[float] cutPoints(vector[float]&, vector[int]&)
 
 cdef class CFImdlp:
     cdef CPPFImdlp *thisptr
-    def __cinit__(self):
-        self.thisptr = new CPPFImdlp()
+    def __cinit__(self, precision=6, debug=False):
+        self.thisptr = new CPPFImdlp(precision, debug)
     def __dealloc__(self):
         del self.thisptr
     def cut_points(self, X, y):

fimdlp/mdlp.py

@@ -19,9 +19,6 @@ class FImdlp(TransformerMixin, BaseEstimator):
         The number of features of the data passed to :meth:`fit`.
     """
 
-    def __init__(self):
-        pass
-
     def _check_params_fit(self, X, y, expected_args, kwargs):
         """Check the common parameters passed to fit"""
         # Check that X and y have correct shape
@@ -93,7 +90,7 @@ class FImdlp(TransformerMixin, BaseEstimator):
         # during fit.
         if X.shape[1] != self.n_features_:
             raise ValueError(
-                "Shape of input is different from what was seen" "in `fit`"
+                "Shape of input is different from what was seen in `fit`"
             )
         print("Cut points for each feature in Iris dataset:")
         yz = self.y_.copy()

sample.py

@@ -6,8 +6,8 @@ data = load_iris()
 X = data.data
 y = data.target
 features = data.feature_names
-# test = FImdlp()
+test = FImdlp()
 # Xcutpoints = test.fit(X, y, features=features).transform(X)
-clf = CFImdlp()
+clf = CFImdlp(debug=True)
 print("Cut points for feature 0 in Iris dataset:")
 print(clf.cut_points(X[:, 0], y))