sdmetrics 0.30.1.dev0__tar.gz → 0.31.1.dev0__tar.gz

{sdmetrics-0.30.1.dev0/sdmetrics.egg-info → sdmetrics-0.31.1.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdmetrics
-Version: 0.30.1.dev0
+Version: 0.31.1.dev0
 Summary: Metrics for Synthetic Data Generation Projects
 Author-email: MIT Data To AI Lab <dailabmit@gmail.com>
 License-Expression: MIT

{sdmetrics-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/pyproject.toml

@@ -148,7 +148,7 @@ convention = 'google'
 add-ignore = ['D107', 'D407', 'D417']
 
 [tool.bumpversion]
-current_version = "0.30.1.dev0"
+current_version = "0.31.1.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-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/sdmetrics/__init__.py

@@ -4,7 +4,7 @@
 
 __author__ = 'MIT Data To AI Lab'
 __email__ = 'dailabmit@gmail.com'
-__version__ = '0.30.1.dev0'
+__version__ = '0.31.1.dev0'
 
 import sys
 import warnings as python_warnings

{sdmetrics-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/sdmetrics/demos/timeseries/metadata.json

@@ -18,7 +18,8 @@
                     "computer_representation": "Float"
                 },
                 "date": {
-                    "sdtype": "datetime"
+                    "sdtype": "datetime",
+                    "datetime_format": "%Y-%m-%d"
                 },
                 "day_of_week": {
                     "sdtype": "numerical",

{sdmetrics-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/sdmetrics/demos.py

@@ -6,15 +6,6 @@ import pathlib
 import pandas as pd
 
 
-def _load_table(metadata, path):
-    datetime_columns = []
-    for column, column_meta in metadata['columns'].items():
-        if column_meta['sdtype'] == 'datetime':
-            datetime_columns.append(column)
-
-    return pd.read_csv(path, parse_dates=datetime_columns)
-
-
 def load_demo(modality='multi_table'):
     """Load demo data of the indicated data modality.
 
@@ -38,8 +29,8 @@ def load_demo(modality='multi_table'):
     real_data = {}
     synthetic_data = {}
     for table, table_meta in metadata['tables'].items():
-        real_data[table] = _load_table(table_meta, demo_path / f'{table}_real.csv')
-        synthetic_data[table] = _load_table(table_meta, demo_path / f'{table}_synthetic.csv')
+        real_data[table] = pd.read_csv(demo_path / f'{table}_real.csv')
+        synthetic_data[table] = pd.read_csv(demo_path / f'{table}_synthetic.csv')
 
     return real_data, synthetic_data, metadata
 

{sdmetrics-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/sdmetrics/multi_table/__init__.py

@@ -32,6 +32,7 @@ from sdmetrics.multi_table.statistical.cardinality_shape_similarity import (
 from sdmetrics.multi_table.statistical.cardinality_statistic_similarity import (
     CardinalityStatisticSimilarity,
 )
+from sdmetrics.multi_table.statistical.constraint_adherence import ConstraintAdherence
 
 __all__ = [
     'detection',
@@ -59,4 +60,5 @@ __all__ = [
     'TVComplement',
     'RangeCoverage',
     'NewRowSynthesis',
+    'ConstraintAdherence',
 ]

{sdmetrics-0.30.1.dev0 → sdmetrics-0.31.1.dev0}/sdmetrics/multi_table/statistical/__init__.py

@@ -6,5 +6,6 @@ from sdmetrics.multi_table.statistical.cardinality_shape_similarity import (
 from sdmetrics.multi_table.statistical.cardinality_statistic_similarity import (
     CardinalityStatisticSimilarity,
 )
+from sdmetrics.multi_table.statistical.constraint_adherence import ConstraintAdherence
 
-__all__ = ['CardinalityShapeSimilarity', 'CardinalityStatisticSimilarity']
+__all__ = ['CardinalityShapeSimilarity', 'CardinalityStatisticSimilarity', 'ConstraintAdherence']

sdmetrics-0.31.1.dev0/sdmetrics/multi_table/statistical/constraint_adherence.py

@@ -0,0 +1,79 @@
+"""Constraint Adherence metric."""
+
+import logging
+import warnings
+
+import numpy as np
+import pandas as pd
+
+from sdmetrics.goal import Goal
+from sdmetrics.multi_table.base import MultiTableMetric
+from sdmetrics.multi_table.statistical.constraints.base import BaseConstraint
+from sdmetrics.multi_table.statistical.constraints.error import ConstraintNotApplicableError
+
+LOGGER = logging.getLogger(__name__)
+
+
+class ConstraintAdherence(MultiTableMetric):
+    """Constraint Adherence metric.
+
+    Compute the fraction of data points in the synthetic data that
+    follow the specified constraint.
+
+    Attributes:
+        name (str):
+            Name to use when reports about this metric are printed.
+        goal (sdmetrics.goal.Goal):
+            The goal of this metric.
+        min_value (float):
+            Minimum value that this metric can take.
+        max_value (float):
+            Maximum value that this metric can take.
+    """
+
+    name = 'ConstraintAdherence'
+    goal = Goal.MAXIMIZE
+    min_value = 0.0
+    max_value = 1.0
+
+    @classmethod
+    def compute(cls, real_data, synthetic_data, metadata, constraint):
+        """Compute the percentage of rows that adhere to the constraint.
+
+        Args:
+            real_data (dict[str, pandas.DataFrame]):
+                The tables from the real dataset, passed as a dictionary of
+                table names and pandas.DataFrames.
+            synthetic_data (dict[str, pandas.DataFrame]):
+                The tables from the synthetic dataset, passed as a dictionary of
+                table names and pandas.DataFrames.
+            metadata (dict):
+                Multi-table metadata dict.
+            constraint (dict):
+                A dictionary that defines the type of constraint and its parameters.
+
+        Returns:
+            float:
+                The proportion of data points in the synthetic data that match
+                the specified constraint format.
+        """
+        constraint = BaseConstraint.load_constraint_from_dict(constraint)
+
+        try:
+            constraint.fit(real_data, metadata)
+            real_score = constraint.get_score(real_data, metadata)
+        except ConstraintNotApplicableError as error:
+            LOGGER.warning(f'Unable to check the constraint against the real data: {error}')
+            real_score = np.nan
+
+        if not pd.isna(real_score) and real_score < 1.0:
+            warnings.warn(
+                f"The real data does not adhere to the '{constraint.__class__.__name__}' "
+                f'constraint ({round(real_score * 100, 2)}% of the rows are valid). '
+            )
+
+        try:
+            return constraint.get_score(synthetic_data, metadata)
+        except ConstraintNotApplicableError as error:
+            LOGGER.warning(f'Unable to check the constraint against the synthetic data: {error}')
+            return np.nan

sdmetrics-0.31.1.dev0/sdmetrics/multi_table/statistical/constraints/__init__.py

@@ -0,0 +1,55 @@
+"""Constraints."""
+
+from sdmetrics.multi_table.statistical.constraints.base import BaseConstraint
+from sdmetrics.multi_table.statistical.constraints.carry_over_columns import CarryOverColumns
+from sdmetrics.multi_table.statistical.constraints.column_formula import ColumnFormula
+from sdmetrics.multi_table.statistical.constraints.denormalized_table import (
+    DenormalizedTable,
+)
+from sdmetrics.multi_table.statistical.constraints.chained_inequality import ChainedInequality
+from sdmetrics.multi_table.statistical.constraints.inequality import Inequality
+from sdmetrics.multi_table.statistical.constraints.fixed_combinations import FixedCombinations
+from sdmetrics.multi_table.statistical.constraints.fixed_increments import FixedIncrements
+from sdmetrics.multi_table.statistical.constraints.fixed_null_combinations import (
+    FixedNullCombinations,
+)
+from sdmetrics.multi_table.statistical.constraints.foreign_to_foreign_key import (
+    ForeignToForeignKey,
+)
+from sdmetrics.multi_table.statistical.constraints.foreign_to_primary_key_subset import (
+    ForeignToPrimaryKeySubset,
+)
+from sdmetrics.multi_table.statistical.constraints.mixed_scales import MixedScales
+from sdmetrics.multi_table.statistical.constraints.polymorphic_relationship import (
+    PolymorphicRelationship,
+)
+from sdmetrics.multi_table.statistical.constraints.primary_to_primary_key_subset import (
+    PrimaryToPrimaryKeySubset,
+)
+from sdmetrics.multi_table.statistical.constraints.reference_table import ReferenceTable
+from sdmetrics.multi_table.statistical.constraints.one_hot_encoding import OneHotEncoding
+from sdmetrics.multi_table.statistical.constraints.range import Range
+from sdmetrics.multi_table.statistical.constraints.referential_hierarchy import (
+    SelfReferentialHierarchy,
+)
+
+__all__ = (
+    BaseConstraint,
+    CarryOverColumns,
+    ColumnFormula,
+    DenormalizedTable,
+    ChainedInequality,
+    Inequality,
+    FixedIncrements,
+    FixedCombinations,
+    FixedNullCombinations,
+    ForeignToForeignKey,
+    ForeignToPrimaryKeySubset,
+    MixedScales,
+    PolymorphicRelationship,
+    PrimaryToPrimaryKeySubset,
+    ReferenceTable,
+    OneHotEncoding,
+    Range,
+    SelfReferentialHierarchy,
+)