working around weird bug with broadcasting over strided arrays

pybasicbayes/distributions/regression.py

@@ -985,8 +985,8 @@ class RobustRegression(Regression):
 
             n, D = y.shape
 
-            scaled_x = x * precisions[:, None]
-            scaled_y = y * precisions[:, None]
+            scaled_x = x * precisions[:, na]
+            scaled_y = y * precisions[:, na]
             xxT = scaled_x.T.dot(x)
             yxT = scaled_y.T.dot(x)
             yyT = scaled_y.T.dot(y)
@@ -1006,7 +1006,9 @@ class RobustRegression(Regression):
             precisions = precisions[~bad]
             n, D = data.shape[0], self.D_out
 
-            scaled_data = data * precisions[: None]
+            # This tile call is suboptimal but without it we can hit issues
+            # with strided data, as in autoregressive models.
+            scaled_data = data * np.tile(precisions[:,None], (1, data.shape[1]))
             statmat = scaled_data.T.dot(data)
 
             xxT, yxT, yyT = \
@@ -1044,7 +1046,7 @@ class RobustRegression(Regression):
             x, y = data
 
         else:
-            x, y = data[:, :self.D_in], data[:, self.D_in:]
+            x, y = data[:, :-self.D_out], data[:, -self.D_out:]
 
         assert x.ndim == y.ndim == 2
         assert x.shape[0] == y.shape[0]
@@ -1111,7 +1113,7 @@ class _ARMixin(object):
         return self.D_out
 
     def predict(self, x):
-        return super(_ARMixin,self).predict(np.atleast_2d(x.ravel()))
+        return super(_ARMixin,self).predict(np.atleast_2d(x))
 
     def rvs(self,lagged_data):
         return super(_ARMixin,self).rvs(