sdv 1.37.4.dev0__tar.gz → 1.37.5.dev0__tar.gz

{sdv-1.37.4.dev0/sdv.egg-info → sdv-1.37.5.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdv
-Version: 1.37.4.dev0
+Version: 1.37.5.dev0
 Summary: Generate synthetic data for single table, multi table and sequential data
 Author-email: "DataCebo, Inc." <info@sdv.dev>
 License-Expression: BUSL-1.1

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/pyproject.toml

@@ -149,7 +149,7 @@ namespaces = false
 version = {attr = 'sdv.__version__'}
 
 [tool.bumpversion]
-current_version = "1.37.4.dev0"
+current_version = "1.37.5.dev0"
 parse = '(?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\.(?P<release>[a-z]+)(?P<candidate>\d+))?'
 serialize = [
     '{major}.{minor}.{patch}.{release}{candidate}',

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/__init__.py

@@ -6,7 +6,7 @@
 
 __author__ = 'DataCebo, Inc.'
 __email__ = 'info@sdv.dev'
-__version__ = '1.37.4.dev0'
+__version__ = '1.37.5.dev0'
 
 
 import sys

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/_utils.py

@@ -417,6 +417,17 @@ def _get_root_tables(relationships):
     return parent_tables - child_tables
 
 
+def _get_max_child_depth(child_map, table):
+    """Return the max child depth for the given table."""
+    max_depth = 1
+    for child in child_map[table]:
+        child_depth = 1 + _get_max_child_depth(child_map, child)
+        if child_depth > max_depth:
+            max_depth = child_depth
+
+    return max_depth
+
+
 def generate_synthesizer_id(synthesizer):
     """Generate a unique identifier for the synthesizer instance.
 

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/evaluation/multi_table.py

@@ -3,8 +3,7 @@
 import warnings
 
 from sdmetrics import visualization
-from sdmetrics.reports.multi_table.diagnostic_report import DiagnosticReport
-from sdmetrics.reports.multi_table.quality_report import QualityReport
+from sdmetrics.reports.multi_table import DiagnosticReport, QualityReport
 
 import sdv.evaluation.single_table as single_table_visualization
 

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/evaluation/single_table.py

@@ -3,8 +3,7 @@
 import warnings
 
 from sdmetrics import visualization
-from sdmetrics.reports.single_table.diagnostic_report import DiagnosticReport
-from sdmetrics.reports.single_table.quality_report import QualityReport
+from sdmetrics.reports import DiagnosticReport, QualityReport
 
 from sdv.errors import VisualizationUnavailableError
 from sdv.evaluation._utils import _prepare_data_visualization
@@ -29,8 +28,9 @@ def _evaluate_quality(real_data, synthetic_data, metadata, verbose=True):
         QualityReport:
             Single table quality report object.
     """
-    if isinstance(metadata, Metadata):
-        metadata = metadata._convert_to_single_table()
+    table_name = metadata._get_single_table_name()
+    real_data = {table_name: real_data}
+    synthetic_data = {table_name: synthetic_data}
 
     quality_report = QualityReport()
     quality_report.generate(real_data, synthetic_data, metadata.to_dict(), verbose)
@@ -56,8 +56,9 @@ def _run_diagnostic(real_data, synthetic_data, metadata, verbose=True):
             Single table diagnostic report object.
     """
     diagnostic_report = DiagnosticReport()
-    if isinstance(metadata, Metadata):
-        metadata = metadata._convert_to_single_table()
+    table_name = metadata._get_single_table_name()
+    real_data = {table_name: real_data}
+    synthetic_data = {table_name: synthetic_data}
 
     diagnostic_report.generate(real_data, synthetic_data, metadata.to_dict(), verbose)
     return diagnostic_report

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/metadata/multi_table.py

@@ -14,6 +14,8 @@ import pandas as pd
 from sdv._utils import (
     _cast_to_iterable,
     _format_invalid_values_string,
+    _get_max_child_depth,
+    _get_root_tables,
     _get_unreferenced_keys,
     _load_data_from_csv,
 )
@@ -289,6 +291,24 @@ class MultiTableMetadata:
         foreign_keys = self._get_all_foreign_keys(table_name)
         return set(foreign_keys).union(self.tables[table_name]._get_primary_and_alternate_keys())
 
+    def _get_max_schema_depth(self):
+        """Calculate the maximum depth of this schema.
+
+        This method traverses all relationships and returns the length of the longest relationship
+        chain between tables.
+
+        Returns:
+            int:
+                The maximum depth of the schema.
+        """
+        max_depth = 1
+        child_map = self._get_child_map()
+        for root_table in _get_root_tables(self.relationships):
+            root_depth = _get_max_child_depth(child_map, root_table)
+            max_depth = root_depth if root_depth > max_depth else max_depth
+
+        return max_depth
+
     def add_relationship(
         self, parent_table_name, child_table_name, parent_primary_key, child_foreign_key
     ):

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0}/sdv/multi_table/hma.py

@@ -203,7 +203,7 @@ class HMASynthesizer(BaseHierarchicalSampler, BaseMultiTableSynthesizer):
             child_tables.add(relationship['child_table_name'])
         for child_table_name in child_tables:
             self.set_table_parameters(child_table_name, {'default_distribution': 'norm'})
-        self._print_estimate_warning()
+        self._validate_schema_complexity()
 
     def set_table_parameters(self, table_name, table_parameters):
         """Update the table's synthesizer instantiation parameters.
@@ -260,42 +260,20 @@ class HMASynthesizer(BaseHierarchicalSampler, BaseMultiTableSynthesizer):
 
         return distributions
 
-    def _print_estimate_warning(self):
-        total_est_cols = 0
-        metadata_columns = self._get_num_data_columns(self.metadata)
-        print_table = []
-        distributions = self._get_distributions()
-        estimated_columns = self._estimate_num_columns(self.metadata, distributions)
-        for table, est_cols in estimated_columns.items():
-            entry = []
-            entry.append(table)
-            entry.append(sum(metadata_columns[table]))
-            total_est_cols += est_cols
-            entry.append(min(est_cols, PERFORMANCE_ALERT_DISPLAY_CAP))
-            print_table.append(entry)
-
-        if total_est_cols > MAX_NUMBER_OF_COLUMNS:
-            display_total = (
-                f'{PERFORMANCE_ALERT_DISPLAY_CAP}+'
-                if total_est_cols > PERFORMANCE_ALERT_DISPLAY_CAP
-                else f'{total_est_cols}'
-            )
-            self._print(
-                'PerformanceAlert: Using the HMASynthesizer on this metadata '
-                'schema is not recommended. To model this data, HMA will '
-                f'generate a large number of columns. ({display_total} columns)\n\n'
-            )
-            self._print(
-                pd.DataFrame(
-                    print_table, columns=['Table Name', '# Columns in Metadata', 'Est # Columns']
-                ).to_string(index=False)
-                + '\n'
-            )
-            self._print(
-                'We recommend simplifying your metadata schema using '
-                "'sdv.utils.poc.simplify_schema'.\nIf this is not possible, please visit "
-                'datacebo.com and reach out to us for enterprise solutions.\n'
+    def _validate_schema_complexity(self):
+        num_tables = len(self.metadata.tables)
+        schema_depth = self.metadata._get_max_schema_depth()
+
+        if num_tables > 5 or schema_depth > 2:
+            error_msg = (
+                'HMASynthesizer is not designed to handle a schema with more than 5 tables or '
+                'relationship depth greater than 2.\n'
+                'Please use SDV Enterprise to model this schema.\n\n'
+                'SDV Enterprise provides access to synthesizers that can easily scale with the '
+                'amount of data and complexity of your schema.\n\n'
+                'For more information, visit datacebo.com'
             )
+            raise SynthesizerInputError(error_msg)
 
     def preprocess(self, data):
         """Transform the raw data to numerical space.

{sdv-1.37.4.dev0 → sdv-1.37.5.dev0/sdv.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdv
-Version: 1.37.4.dev0
+Version: 1.37.5.dev0
 Summary: Generate synthetic data for single table, multi table and sequential data
 Author-email: "DataCebo, Inc." <info@sdv.dev>
 License-Expression: BUSL-1.1