sdmetrics 0.31.1.dev0__tar.gz → 0.31.2.dev0__tar.gz

{sdmetrics-0.31.1.dev0/sdmetrics.egg-info → sdmetrics-0.31.2.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdmetrics
-Version: 0.31.1.dev0
+Version: 0.31.2.dev0
 Summary: Metrics for Synthetic Data Generation Projects
 Author-email: MIT Data To AI Lab <dailabmit@gmail.com>
 License-Expression: MIT

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/pyproject.toml

@@ -148,7 +148,7 @@ convention = 'google'
 add-ignore = ['D107', 'D407', 'D417']
 
 [tool.bumpversion]
-current_version = "0.31.1.dev0"
+current_version = "0.31.2.dev0"
 parse = '(?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\.(?P<release>[a-z]+)(?P<candidate>\d+))?'
 serialize = [
     '{major}.{minor}.{patch}.{release}{candidate}',

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/__init__.py

@@ -4,7 +4,7 @@
 
 __author__ = 'MIT Data To AI Lab'
 __email__ = 'dailabmit@gmail.com'
-__version__ = '0.31.1.dev0'
+__version__ = '0.31.2.dev0'
 
 import sys
 import warnings as python_warnings

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/reports/base_report.py

@@ -75,7 +75,19 @@ class BaseReport:
         )
         raise ValueError(error_message)
 
-    def _validate(self, real_data, synthetic_data, metadata):
+    def _validate_constraints_input(self, constraints):
+        """Validate that the constraints are ``None`` or a list of dictionaries."""
+        if constraints is None:
+            return
+
+        is_list = isinstance(constraints, list)
+        if not is_list or not all(isinstance(constraint, dict) for constraint in constraints):
+            raise ValueError(
+                f"{self.__class__.__name__} expects 'constraints' parameter to be "
+                "a list of dictionaries, each one with the keys 'class_name' and 'parameters'."
+            )
+
+    def _validate(self, real_data, synthetic_data, metadata, constraints=None):
         """Validate the inputs.
 
         Args:
@@ -85,10 +97,13 @@ class BaseReport:
                 The synthetic data.
             metadata (dict):
                 The metadata of the table.
+            constraints (list[dict] or None):
+                The constraints to evaluate. Defaults to None.
         """
         self._validate_data_format(real_data, synthetic_data)
         _validate_metadata(metadata)
         self._validate_metadata_matches_data(real_data, synthetic_data, metadata)
+        self._validate_constraints_input(constraints)
 
     @staticmethod
     def convert_datetimes(real_data, synthetic_data, metadata):
@@ -130,12 +145,14 @@ class BaseReport:
         if verbose:
             sys.stdout.write(f'Overall Score (Average): {round(self._overall_score * 100, 2)}%\n\n')
 
-    def _get_skipped_properties(self, metadata):
+    def _get_skipped_properties(self, metadata, constraints=None):
         """Return properties that should not be computed for the metadata.
 
         Args:
             metadata (dict):
                 The metadata dict.
+            constraints (list[dict] or None):
+                The constraints given to the report. Defaults to None.
 
         Returns:
             set[str]:
@@ -143,7 +160,7 @@ class BaseReport:
         """
         return set()
 
-    def generate(self, real_data, synthetic_data, metadata, verbose=True):
+    def generate(self, real_data, synthetic_data, metadata, constraints=None, verbose=True):
         """Generate report.
 
         This method generates the report by iterating through each property and calculating
@@ -156,13 +173,16 @@ class BaseReport:
                 The synthetic data.
             metadata (dict):
                 The metadata, which contains each column's data type as well as relationships.
+            constraints (list[dict] or None):
+                A list of constraints to evaluate their adherence, each represented as a
+                dictionary with a ``class_name`` and a ``parameters`` key. Defaults to None.
             verbose (bool):
                 Whether or not to print report summary and progress.
         """
         real_data = deepcopy(real_data)
         synthetic_data = deepcopy(synthetic_data)
-        self._validate(real_data, synthetic_data, metadata)
-        self._skipped_properties = self._get_skipped_properties(metadata)
+        self._validate(real_data, synthetic_data, metadata, constraints)
+        self._skipped_properties = self._get_skipped_properties(metadata, constraints)
         self._original_datetime_columns = self.convert_datetimes(
             real_data, synthetic_data, metadata
         )
@@ -192,15 +212,22 @@ class BaseReport:
                 property_instance.is_computed = False
                 property_instance.details = pd.DataFrame()
                 if verbose:
-                    sys.stdout.write(
-                        f'{property_description}: N/A\n{self._skipped_property_message}\n\n'
+                    skipped_message = getattr(
+                        property_instance, '_skipped_message', self._skipped_property_message
                     )
+                    sys.stdout.write(f'{property_description}: N/A\n{skipped_message}\n\n')
                     sys.stdout.flush()
 
                 continue
 
+            property_arguments = {}
+            if property_name == 'Constraint Validity':
+                property_arguments['constraints'] = constraints
+
             if verbose:
-                num_iterations = int(property_instance._get_num_iterations(metadata))
+                num_iterations = int(
+                    property_instance._get_num_iterations(metadata, **property_arguments)
+                )
                 progress_bar = tqdm.tqdm(
                     total=num_iterations, file=sys.stdout, bar_format='{desc}|{bar}{r_bar}|'
                 )
@@ -220,7 +247,7 @@ class BaseReport:
 
             self._properties[property_name].num_rows_subsample = self.num_rows_subsample
             score = self._properties[property_name].get_score(
-                real_data, synthetic_data, metadata, progress_bar=progress_bar
+                real_data, synthetic_data, metadata, progress_bar=progress_bar, **property_arguments
             )
             scores.append(score)
             if verbose:

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/reports/base_unified_report.py

@@ -16,21 +16,26 @@ class BaseUnifiedReport(BaseMultiTableReport):
     })
     _skipped_property_message = 'This property does not apply to single-table data.'
 
-    def _get_skipped_properties(self, metadata):
+    def _get_skipped_properties(self, metadata, constraints=None):
         """Return properties unavailable to single-table data.
 
         Args:
             metadata (dict):
                 The metadata dict.
+            constraints (list[dict] or None):
+                The constraints given to the report. Defaults to None.
 
         Returns:
             set[str]:
                 Names of properties to skip.
         """
+        skipped_properties = super()._get_skipped_properties(metadata, constraints)
         if len(metadata.get('tables', {})) == 1:
-            return self._SINGLE_TABLE_SKIPPED_PROPERTIES
+            skipped_properties.update(self._SINGLE_TABLE_SKIPPED_PROPERTIES)
+        if not constraints:
+            skipped_properties.add('Constraint Validity')
 
-        return super()._get_skipped_properties(metadata)
+        return skipped_properties
 
     def _validate_data_format(self, real_data, synthetic_data):
         """Validate that the real and synthetic data have compatible formats.
@@ -75,7 +80,7 @@ class BaseUnifiedReport(BaseMultiTableReport):
         super()._validate_property_generated(property_name)
         self._check_property_single_table(property_name)
 
-    def _validate(self, real_data, synthetic_data, metadata):
+    def _validate(self, real_data, synthetic_data, metadata, constraints=None):
         """Validate the inputs.
 
         Args:
@@ -85,6 +90,8 @@ class BaseUnifiedReport(BaseMultiTableReport):
                 The synthetic data.
             metadata (dict):
                 The metadata.
+            constraints (list[dict] or None):
+                The constraints to evaluate. Defaults to None.
         """
         _validate_unified_metadata(metadata)
         self.table_names = list(metadata.get('tables', []))
@@ -94,6 +101,7 @@ class BaseUnifiedReport(BaseMultiTableReport):
             synthetic_data,
             metadata,
         )
+        self._validate_constraints_input(constraints)
 
     def get_properties(self):
         """Return the property score.

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/reports/diagnostic_report.py

@@ -1,7 +1,12 @@
 """Unified diagnostic report."""
 
 from sdmetrics.reports.base_unified_report import BaseUnifiedReport
-from sdmetrics.reports.multi_table._properties import DataValidity, RelationshipValidity, Structure
+from sdmetrics.reports.multi_table._properties import (
+    ConstraintValidity,
+    DataValidity,
+    RelationshipValidity,
+    Structure,
+)
 
 
 class DiagnosticReport(BaseUnifiedReport):
@@ -17,6 +22,7 @@ class DiagnosticReport(BaseUnifiedReport):
             'Data Validity': DataValidity(),
             'Data Structure': Structure(),
             'Relationship Validity': RelationshipValidity(),
+            'Constraint Validity': ConstraintValidity(),
         }
 
     def _validate_metadata_matches_data(self, real_data, synthetic_data, metadata):

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/reports/multi_table/_properties/constraint_validity.py

@@ -17,6 +17,7 @@ class ConstraintValidity(BaseMultiTableProperty):
     """
 
     _num_iteration_case = 'constraint'
+    _skipped_message = 'No constraints were provided.'
 
     def _get_num_iterations(self, metadata, constraints=None):
         """Get the number of iterations for the property, which is one per constraint."""

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0}/sdmetrics/reports/multi_table/base_multi_table_report.py

@@ -84,7 +84,7 @@ class BaseMultiTableReport(BaseReport):
 
         self._validate_relationships(real_data, synthetic_data, metadata)
 
-    def generate(self, real_data, synthetic_data, metadata, verbose=True):
+    def generate(self, real_data, synthetic_data, metadata, constraints=None, verbose=True):
         """Generate report.
 
         This method generates the report by iterating through each property and calculating
@@ -97,10 +97,12 @@ class BaseMultiTableReport(BaseReport):
                 The synthetic data.
             metadata (dict):
                 The metadata, which contains each column's data type as well as relationships.
+            constraints (list[dict] or None):
+                A list of constraints to evaluate their adherence. Defaults to None.
             verbose (bool):
                 Whether or not to print report summary and progress.
         """
-        results = super().generate(real_data, synthetic_data, metadata, verbose)
+        results = super().generate(real_data, synthetic_data, metadata, constraints, verbose)
         self.table_names = list(metadata.get('tables', {}).keys())
 
         return results
@@ -173,6 +175,9 @@ class BaseMultiTableReport(BaseReport):
         if property_name == 'Data Structure':
             return self._properties[property_name].get_visualization(table_name)
 
+        if property_name == 'Constraint Validity':
+            return self._properties[property_name].get_visualization()
+
         if table_name is None:
             raise ValueError('Please provide a table name to get a visualization for the property.')
 

{sdmetrics-0.31.1.dev0 → sdmetrics-0.31.2.dev0/sdmetrics.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdmetrics
-Version: 0.31.1.dev0
+Version: 0.31.2.dev0
 Summary: Metrics for Synthetic Data Generation Projects
 Author-email: MIT Data To AI Lab <dailabmit@gmail.com>
 License-Expression: MIT