Handle nans

Author: eirol

epochutils/stats/distributions.py

@@ -282,15 +282,12 @@ def cdf(self, x, _apply_transform=True):
         if np.isscalar(x):
             return self._cdf_scalar(x)
         else:
-            x = np.atleast_1d(x)
-
-            nan_mask = np.isnan(x)
-            result = np.full_like(x, np.nan)
-            result[~nan_mask] = np.array([self._cdf_scalar(a) for a in x[~nan_mask]])
-
-            return result
+            return np.array([self._cdf_scalar(a) for a in np.atleast_1d(x)])
 
     def _cdf_scalar(self, x):
+        if np.isnan(x):
+            return np.nan
+
         if self.lower_bound is not None and x <= self.lower_bound:
             return 0