Fixing least squares out-transformation bug

tests/test_regression.py

@@ -64,3 +64,24 @@ def test_simulation(orca_session):
 
     assert orca.get_table('obs').to_frame()['a_predicted'].equals(m.predicted_values)
 
+
+def test_out_transform(orca_session):
+    """
+    Test transformation of the predicted values.
+    
+    """
+    modelmanager.initialize()
+
+    m = OLSRegressionStep()
+    m.tables = 'obs'
+    m.model_expression = 'a ~ b'
+    m.fit()
+
+    m.out_column = 'a_predicted'
+    m.out_transform = 'np.exp'
+    m.run()
+
+    predictions = m.predicted_values.apply(np.exp)
+
+    assert orca.get_table('obs').to_frame()['a_predicted'].equals(predictions)
+

urbansim_templates/models/regression.py

@@ -1,5 +1,6 @@
 from __future__ import print_function
 
+import math
 import numpy as np
 import pandas as pd
 from datetime import datetime as dt
@@ -69,10 +70,12 @@ class OLSRegressionStep(TemplateStep):
         side variable from the model expression will be used. Replaces the `out_fname` 
         argument in UrbanSim.
     
-    out_transform : callable, optional
-        Transformation to apply to the predicted values, for example to reverse a 
-        transformation of the left-hand-side variable in the model expression. Replaces
-        the `ytransform` argument in UrbanSim.
+    out_transform : str, optional
+        Element-wise transformation to apply to the predicted values, for example to
+        reverse a transformation of the left-hand-side variable in the model expression.
+        This should be provided as a string containing a function name. Supports anything
+        from NumPy or Python's built-in math library, for example 'np.exp' or
+        'math.floor'. Replaces the `ytransform` argument in UrbanSim.
     
     out_filters : str or list of str, optional
         Filters to apply to the data before simulation. If not provided, no filters will
@@ -168,7 +171,7 @@ def fit(self):
         """
         self.model = RegressionModel(model_expression=self.model_expression,
                 fit_filters=self.filters, predict_filters=self.out_filters,
-                ytransform=self.out_transform, name=self.name)
+                ytransform=None, name=self.name)
 
         df = get_data(tables = self.tables,
                       filters = self.filters,
@@ -207,7 +210,7 @@ def run(self):
         self.predicted_values = values
 
         if self.out_transform is not None:
-            values = self.out_transform(values)
+            values = values.apply(eval(self.out_transform))
 
         colname = self._get_out_column()
         tabname = self._get_out_table()