Benchmark Supported Data Types

.github/workflows/integration.yml

@@ -29,4 +29,9 @@ jobs:
           python -m pip install --upgrade pip
           python -m pip install invoke .[test]
     - name: Run integration tests
-      run: invoke integration
+      env:
+        PYDRIVE_CREDENTIALS: ${{ secrets.PYDRIVE_CREDENTIALS }}
+
+      run: |
+        invoke integration
+        invoke benchmark-dtypes

pyproject.toml

@@ -62,6 +62,7 @@ test = [
     'rundoc>=0.4.3,<0.5',
     'pytest-runner >= 2.11.1',
     'tomli>=2.0.0,<3',
+    'pydrive',
 ]
 pomegranate = ['pomegranate>=0.14.3,<0.15']
 dev = [

sdv/_utils.py

@@ -9,6 +9,7 @@
 from pathlib import Path
 
 import pandas as pd
+from pandas.api.types import is_float, is_integer
 from pandas.core.tools.datetimes import _guess_datetime_format_for_array
 from rdt.transformers.utils import _GENERATORS
 
@@ -439,3 +440,11 @@ def get_possible_chars(regex, num_subpatterns=None):
         possible_chars += _get_chars_for_option(option, params)
 
     return possible_chars
+
+
+def _is_numerical(value):
+    """Determine if the input is a numerical type or not."""
+    try:
+        return is_integer(value) or is_float(value)
+    except Exception:
+        return False

sdv/constraints/tabular.py

@@ -36,7 +36,7 @@
 import numpy as np
 import pandas as pd
 
-from sdv._utils import _convert_to_timedelta, _create_unique_name, _is_datetime_type
+from sdv._utils import _convert_to_timedelta, _create_unique_name, _is_datetime_type, _is_numerical
 from sdv.constraints.base import Constraint
 from sdv.constraints.errors import (
     AggregateConstraintsError,
@@ -604,7 +604,7 @@ def _validate_metadata_specific_to_constraint(metadata, **kwargs):
         sdtype = metadata.columns.get(column_name, {}).get('sdtype')
         value = kwargs.get('value')
         if sdtype == 'numerical':
-            if not isinstance(value, (int, float)):
+            if not _is_numerical(value):
                 raise ConstraintMetadataError("'value' must be an int or float.")
 
         elif sdtype == 'datetime':
@@ -632,7 +632,7 @@ def _validate_init_inputs(column_name, value, relation):
         if relation not in ['>', '>=', '<', '<=']:
             raise ValueError('`relation` must be one of the following: `>`, `>=`, `<`, `<=`')
 
-        if not (isinstance(value, (int, float)) or value_is_datetime):
+        if not (_is_numerical(value) or value_is_datetime):
             raise ValueError('`value` must be a number or a string that represents a datetime.')
 
         if value_is_datetime and not isinstance(value, str):
@@ -1071,9 +1071,7 @@ def _validate_init_inputs(low_value, high_value):
         if values_are_datetimes and not values_are_strings:
             raise ValueError('Datetime must be represented as a string.')
 
-        values_are_numerical = bool(
-            isinstance(low_value, (int, float)) and isinstance(high_value, (int, float))
-        )
+        values_are_numerical = bool(_is_numerical(low_value) and _is_numerical(high_value))
         if not (values_are_numerical or values_are_datetimes):
             raise ValueError(
                 '``low_value`` and ``high_value`` must be a number or a string that '
@@ -1092,7 +1090,7 @@ def _validate_metadata_specific_to_constraint(metadata, **kwargs):
         high_value = kwargs.get('high_value')
         low_value = kwargs.get('low_value')
         if sdtype == 'numerical':
-            if not isinstance(high_value, (int, float)) or not isinstance(low_value, (int, float)):
+            if not _is_numerical(high_value) or not _is_numerical(low_value):
                 raise ConstraintMetadataError(
                     "Both 'high_value' and 'low_value' must be ints or floats"
                 )

tasks.py

@@ -34,6 +34,11 @@ def integration(c):
     c.run('python -m pytest ./tests/integration --reruns 3')
 
 
+@task
+def benchmark_dtypes(c):
+    c.run('python -m pytest ./tests/benchmark/supported_dtypes_benchmark.py')
+
+
 def _get_minimum_versions(dependencies, python_version):
     min_versions = {}
     for dependency in dependencies:

tests/benchmark/__init__.py

@@ -0,0 +1 @@
+"""SDV benchmarking module."""