Merge pull request #55 from cogito233/main

Fix NaN problem while normailze the data

causallearn/utils/KCI/KCI.py

@@ -146,7 +146,11 @@ def kernel_matrix(self, data_x, data_y):
             raise Exception('Undefined kernel function')
 
         data_x = stats.zscore(data_x, axis=0)
+        data_x[np.isnan(data_x)] = 0.
+
         data_y = stats.zscore(data_y, axis=0)
+        data_y[np.isnan(data_y)] = 0.
+
         Kx = kernelX.kernel(data_x)
         Ky = kernelY.kernel(data_y)
         return Kx, Ky
@@ -323,8 +327,13 @@ def kernel_matrix(self, data_x, data_y, data_z):
         """
         # normalize the data
         data_x = stats.zscore(data_x, axis=0)
+        data_x[np.isnan(data_x)] = 0.
+
         data_y = stats.zscore(data_y, axis=0)
+        data_y[np.isnan(data_y)] = 0.
+
         data_z = stats.zscore(data_z, axis=0)
+        data_z[np.isnan(data_z)] = 0.
 
         # concatenate x and z
         data_x = np.concatenate((data_x, 0.5 * data_z), axis=1)

causallearn/utils/cit.py

@@ -1,4 +1,5 @@
 from math import log, sqrt
+from collections.abc import Iterable
 
 import numpy as np
 from scipy.stats import chi2, norm
@@ -59,9 +60,18 @@ def _unique(column):
         }
 
     def kci(self, X, Y, condition_set):
+        if type(X) == int:
+            X = [X]
+        elif type(X) != list:
+            Y = list(X)
+        if type(Y) == int:
+            Y = [Y]
+        elif type(Y) != list:
+            Y = list(Y)
+
         if len(condition_set) == 0:
-            return self.kci_ui.compute_pvalue(self.data[:, [X]], self.data[:, [Y]])[0]
-        return self.kci_ci.compute_pvalue(self.data[:, [X]], self.data[:, [Y]], self.data[:, list(condition_set)])[0]
+            return self.kci_ui.compute_pvalue(self.data[:, X], self.data[:, Y])[0]
+        return self.kci_ci.compute_pvalue(self.data[:, X], self.data[:, Y], self.data[:, list(condition_set)])[0]
 
     def fisherz(self, X, Y, condition_set):
         """
@@ -326,7 +336,23 @@ def __call__(self, X, Y, condition_set=None, *args):
         else:
             assert len(args) == 2, "Arguments other than skel and prt_m are provided for mc_fisherz."
         if condition_set is None: condition_set = tuple()
-        assert X not in condition_set and Y not in condition_set, "X, Y cannot be in condition_set."
+
+        if type(X) == int and type(Y) == int:
+            assert X not in condition_set and Y not in condition_set, "X, Y cannot be in condition_set."
+        else:
+            if isinstance(X, Iterable):
+                assert len(set(condition_set).intersection(X)) == 0, "X cannot be in condition_set."
+            elif isinstance(X, int):
+                assert X not in condition_set, "X cannot be in condition_set."
+            else:
+                raise Exception("Undefined type of X, X should be int or Iterable")
+            if isinstance(Y, Iterable):
+                assert len(set(condition_set).intersection(Y)) == 0, "Y cannot be in condition_set."
+            elif isinstance(Y, int):
+                assert Y not in condition_set, "Y cannot be in condition_set."
+            else:
+                raise Exception("Undefined type of Y, Y should be int or Iterable")
+
         i, j = (X, Y) if (X < Y) else (Y, X)
         cache_key = (i, j, frozenset(condition_set))