Merge pull request #99 from UDST/v1-patch

[0.1.2] Small MNL patch

CHANGELOG.md

@@ -1,5 +1,9 @@
 # UrbanSim Templates change log
 
+### 0.1.2 (2019-02-28)
+
+- fixes a crash in small MNL simulation
+
 ### 0.1.1 (2019-02-05)
 
 - production release

docs/source/index.rst

@@ -10,7 +10,7 @@ UrbanSim Templates provides building blocks for Orca-based simulation models. It
 
 The library contains templates for common types of model steps, plus a tool called ModelManager that runs as an extension to the `Orca <https://udst.github.io/orca>`__ task orchestrator. ModelManager can register template-based model steps with the orchestrator, save them to disk, and automatically reload them for future sessions.
 
-v0.1.1, released February 5, 2019
+v0.1.2, released February 28, 2019
 
 
 Contents

setup.py

@@ -6,7 +6,7 @@
 
 setup(
     name='urbansim_templates',
-    version='0.1.1',
+    version='0.1.2',
     description='UrbanSim extension for managing model steps',
     author='UrbanSim Inc.',
     author_email='info@urbansim.com',

tests/test_small_multinomial_logit.py

@@ -11,12 +11,22 @@
 
 @pytest.fixture
 def orca_session():
-    d1 = {'a': np.random.random(100),
-          'b': np.random.random(100),
+    d1 = {'id': np.arange(100),
+          'building_id': np.arange(100),
+          'a': np.random.random(100),
           'choice': np.random.randint(3, size=100)}
+
+    d2 = {'building_id': np.arange(100),
+          'b': np.random.random(100)}
+
+    households = pd.DataFrame(d1).set_index('id')
+    orca.add_table('households', households)
+
+    buildings = pd.DataFrame(d2).set_index('building_id')
+    orca.add_table('buildings', buildings)
 
-    obs = pd.DataFrame(d1)
-    orca.add_table('obs', obs)
+    orca.broadcast(cast='buildings', onto='households', 
+                   cast_index=True, onto_on='building_id')
 
 
 def test_template_validity():
@@ -35,7 +45,7 @@ def test_small_mnl(orca_session):
     modelmanager.initialize()
 
     m = SmallMultinomialLogitStep()
-    m.tables = 'obs'
+    m.tables = ['households', 'buildings']
     m.choice_column = 'choice'
     m.model_expression = OrderedDict([
             ('intercept', [1,2]), ('a', [0,2]), ('b', [0,2])])
@@ -51,6 +61,12 @@ def test_small_mnl(orca_session):
 
     print(m.model_expression)
 
+    # TEST SIMULATION
+    m.out_column = 'simulated_choice'
+
+    m.run()
+    print(orca.get_table('households').to_frame())
+
     modelmanager.initialize()
     m = modelmanager.get_step('small-mnl-test')
     assert(m.model_expression is not None)

urbansim_templates/__init__.py

@@ -1 +1 @@
-version = __version__ = '0.1.1'
+version = __version__ = '0.1.2'

urbansim_templates/models/shared.py

@@ -168,16 +168,6 @@ def _get_data(self, task='fit'):
         if isinstance(self.model_expression, str):
             expr_cols = util.columns_in_formula(self.model_expression)
 
-        # This is for PyLogit model expressions
-        elif isinstance(self.model_expression, OrderedDict):
-            # TO DO - check that this works in Python 2.7
-            expr_cols = [t[0] for t in list(self.model_expression.items()) \
-                         if t[0] != 'intercept']
-            # TO DO - not very general, maybe we should just override the method
-            # TO DO - and this only applies to the fit condition
-            if self.choice_column is not None:
-                expr_cols += [self.choice_column]
-
         if (task == 'fit'):
             tables = self.tables
             columns = expr_cols + util.columns_in_filters(self.filters)
@@ -204,16 +194,6 @@ def _get_data(self, task='fit'):
         return df
 
 
-    def _get_filter_columns(self):
-        """
-        THIS METHOD DOES NOT WORK YET.   
-        
-        Return list of column names referenced in the filters.
-        
-        """
-        return
-
-
     def _get_out_column(self):
         """
         Return name of the column to save data to. This is 'out_column' if it exsits,

urbansim_templates/models/small_multinomial_logit.py

@@ -10,9 +10,9 @@
 from choicemodels import MultinomialLogit
 import orca
 
-from ..utils import update_column
-from .. import modelmanager
-from .shared import TemplateStep
+from urbansim_templates import modelmanager
+from urbansim_templates.models import TemplateStep
+from urbansim_templates.utils import get_data, update_column
 
 
 @modelmanager.template
@@ -87,8 +87,6 @@ class SmallMultinomialLogitStep(TemplateStep):
         in the primary output table, it will be created. If not provided, the 
         `choice_column` will be used. Replaces the `out_fname` argument in UrbanSim.
         
-        # TO DO - auto-generation not yet working; column must exist in the primary table
-
     out_filters : str or list of str, optional
         Filters to apply to the data before simulation. If not provided, no filters will
         be applied. Replaces the `predict_filters` argument in UrbanSim.
@@ -312,14 +310,22 @@ def fit(self):
         with Orca or ModelManager until the `register()` method is run. 
         
         """
-        long_df = self._to_long(self._get_data())
+        expr_cols = [t[0] for t in list(self.model_expression.items()) \
+                     if t[0] != 'intercept']
+
+        df = get_data(tables = self.tables,
+                      filters = self.filters,
+                      extra_columns = expr_cols + [self.choice_column])
+
+        long_df = self._to_long(df)
 
         # Set initial coefs to 0 if none provided
         pc = self._get_param_count()
         if (self.initial_coefs is None) or (len(self.initial_coefs) != pc):
             self.initial_coefs = np.zeros(pc).tolist()
 
-        model = MultinomialLogit(data=long_df, observation_id_col='_obs_id',
+        model = MultinomialLogit(data=long_df, 
+                                 observation_id_col='_obs_id',
                                  choice_col='_chosen',
                                  model_expression=self.model_expression,
                                  model_labels=self.model_labels,
@@ -352,7 +358,14 @@ def run(self):
         model step.
         
         """
-        df = self._get_data('predict')
+        expr_cols = [t[0] for t in list(self.model_expression.items()) \
+                     if t[0] != 'intercept']
+
+        df = get_data(tables = self.out_tables,
+                      fallback_tables = self.tables,
+                      filters = self.out_filters,
+                      extra_columns = expr_cols)
+
         long_df = self._to_long(df, 'predict')
 
         num_obs = len(df)