Add documentation for and invert default

sdv/constraints/base.py

@@ -4,6 +4,7 @@
 import importlib
 import inspect
 import logging
+import warnings
 
 import pandas as pd
 from copulas.multivariate.gaussian import GaussianMultivariate
@@ -111,7 +112,7 @@ class Constraint(metaclass=ConstraintMeta):
     def _identity(self, table_data):
         return table_data
 
-    def __init__(self, handling_strategy, fit_columns_model=True):
+    def __init__(self, handling_strategy, fit_columns_model=False):
         self.fit_columns_model = fit_columns_model
         if handling_strategy == 'transform':
             self.filter_valid = self._identity
@@ -226,6 +227,14 @@ def _validate_constraint_columns(self, table_data):
         """
         missing_columns = [col for col in self.constraint_columns if col not in table_data.columns]
         if missing_columns:
+            if not self._columns_model:
+                warning_message = (
+                    'The following constraint columns are missing from the table data '
+                    f'- `{missing_columns}`. When `fit_columns_model` is `False` and one or more '
+                    'constraint columns are missing, reject sampling can become slow.'
+                )
+                warnings.warn(warning_message, UserWarning)
+
             all_columns_missing = len(missing_columns) == len(self.constraint_columns)
             if self._columns_model is None or all_columns_missing:
                 raise MissingConstraintColumnError()

sdv/constraints/tabular.py

@@ -80,14 +80,18 @@ class UniqueCombinations(Constraint):
         handling_strategy (str):
             How this Constraint should be handled, which can be ``transform``,
             ``reject_sampling`` or ``all``. Defaults to ``transform``.
+        fit_columns_model (bool):
+            If False, reject sampling will be used to handle conditional sampling.
+            Otherwise, a model will be trained and used to sample other columns
+            based on the conditioned column. Defaults to False.
     """
 
     _separator = None
     _joint_column = None
     _combinations_to_uuids = None
     _uuids_to_combinations = None
 
-    def __init__(self, columns, handling_strategy='transform', fit_columns_model=True):
+    def __init__(self, columns, handling_strategy='transform', fit_columns_model=False):
         if len(columns) < 2:
             raise ValueError('UniqueCombinations requires at least two constraint columns.')
 
@@ -214,6 +218,10 @@ class GreaterThan(Constraint):
         handling_strategy (str):
             How this Constraint should be handled, which can be ``transform``
             or ``reject_sampling``. Defaults to ``transform``.
+        fit_columns_model (bool):
+            If False, reject sampling will be used to handle conditional sampling.
+            Otherwise, a model will be trained and used to sample other columns
+            based on the conditioned column. Defaults to False.
         drop (str):
             Which column to drop during transformation. Can be ``'high'``,
             ``'low'`` or ``None``.
@@ -290,7 +298,7 @@ def _get_columns_to_reconstruct(self):
         return column
 
     def __init__(self, low, high, strict=False, handling_strategy='transform',
-                 fit_columns_model=True, drop=None, scalar=None):
+                 fit_columns_model=False, drop=None, scalar=None):
         self._strict = strict
         self._drop = drop
         self._scalar = scalar
@@ -485,12 +493,16 @@ class Positive(GreaterThan):
         handling_strategy (str):
             How this Constraint should be handled, which can be ``transform``
             or ``reject_sampling``. Defaults to ``transform``.
+        fit_columns_model (bool):
+            If False, reject sampling will be used to handle conditional sampling.
+            Otherwise, a model will be trained and used to sample other columns
+            based on the conditioned column. Defaults to False.
         drop (bool):
             Whether to drop columns during transformation.
     """
 
     def __init__(self, columns, strict=False, handling_strategy='transform',
-                 fit_columns_model=True, drop=False):
+                 fit_columns_model=False, drop=False):
         drop = 'high' if drop else None
         super().__init__(handling_strategy=handling_strategy,
                          fit_columns_model=fit_columns_model,
@@ -515,12 +527,16 @@ class Negative(GreaterThan):
         handling_strategy (str):
             How this Constraint should be handled, which can be ``transform``
             or ``reject_sampling``. Defaults to ``transform``.
+        fit_columns_model (bool):
+            If False, reject sampling will be used to handle conditional sampling.
+            Otherwise, a model will be trained and used to sample other columns
+            based on the conditioned column. Defaults to False.
         drop (bool):
             Whether to drop columns during transformation.
     """
 
     def __init__(self, columns, strict=False, handling_strategy='transform',
-                 fit_columns_model=True, drop=False):
+                 fit_columns_model=False, drop=False):
         drop = 'low' if drop else None
         super().__init__(handling_strategy=handling_strategy,
                          fit_columns_model=fit_columns_model,
@@ -633,6 +649,10 @@ class Between(Constraint):
         handling_strategy (str):
             How this Constraint should be handled, which can be ``transform``
             or ``reject_sampling``. Defaults to ``transform``.
+        fit_columns_model (bool):
+            If False, reject sampling will be used to handle conditional sampling.
+            Otherwise, a model will be trained and used to sample other columns
+            based on the conditioned column. Defaults to False.
         high_is_scalar(bool or None):
             Whether or not the value for high is a scalar or a column name.
             If ``None``, this will be determined during the ``fit`` method
@@ -646,7 +666,7 @@ class Between(Constraint):
     _transformed_column = None
 
     def __init__(self, column, low, high, strict=False, handling_strategy='transform',
-                 fit_columns_model=True, high_is_scalar=None, low_is_scalar=None):
+                 fit_columns_model=False, high_is_scalar=None, low_is_scalar=None):
         self.constraint_column = column
         self.constraint_columns = (column,)
         self._low = low

tests/integration/tabular/test_base.py

@@ -180,7 +180,8 @@ def test_conditional_sampling_constraint_uses_columns_model(gm_mock):
     # Setup
     constraint = UniqueCombinations(
         columns=['city', 'state'],
-        handling_strategy='transform'
+        handling_strategy='transform',
+        fit_columns_model=True,
     )
     data = pd.DataFrame({
         'city': ['LA', 'SF', 'CHI', 'LA', 'LA'],

tests/unit/constraints/test_base.py

@@ -260,7 +260,7 @@ def test_fit_trains_column_model(self, ht_mock, gm_mock):
             'a': [1, 2, 3],
             'b': [4, 5, 6]
         })
-        instance = Constraint(handling_strategy='transform')
+        instance = Constraint(handling_strategy='transform', fit_columns_model=True)
         instance.constraint_columns = ('a', 'b')
 
         # Run