sdgym 0.15.2.dev0__tar.gz → 0.16.1.dev0__tar.gz

{sdgym-0.15.2.dev0/sdgym.egg-info → sdgym-0.16.1.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdgym
-Version: 0.15.2.dev0
+Version: 0.16.1.dev0
 Summary: Benchmark tabular synthetic data generators using a variety of datasets
 Author-email: "DataCebo, Inc." <info@sdv.dev>
 License-Expression: BUSL-1.1
@@ -63,8 +63,8 @@ Requires-Dist: torch>=2.6.0; python_version >= "3.13" and python_version < "3.14
 Requires-Dist: torch>=2.9.0; python_version >= "3.14"
 Requires-Dist: tqdm>=4.66.3
 Requires-Dist: XlsxWriter>=1.2.8
-Requires-Dist: rdt>=1.18.2; python_version < "3.14"
-Requires-Dist: rdt>=1.20.0; python_version >= "3.14"
+Requires-Dist: rdt<2,>=1.18.2; python_version < "3.14"
+Requires-Dist: rdt<2,>=1.20.0; python_version >= "3.14"
 Requires-Dist: sdmetrics>=0.29.0
 Requires-Dist: sdv<2,>=1.37.0
 Provides-Extra: dask

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/pyproject.toml

@@ -61,8 +61,8 @@ dependencies = [
     "torch>=2.9.0;python_version>='3.14'",
     'tqdm>=4.66.3',
     'XlsxWriter>=1.2.8',
-    "rdt>=1.18.2;python_version<'3.14'",
-    "rdt>=1.20.0;python_version>='3.14'",
+    "rdt>=1.18.2,<2;python_version<'3.14'",
+    "rdt>=1.20.0,<2;python_version>='3.14'",
     "sdmetrics>=0.29.0",
     "sdv>=1.37.0,<2",
 ]
@@ -177,7 +177,7 @@ namespaces = false
 version = {attr = 'sdgym.__version__'}
 
 [tool.bumpversion]
-current_version = "0.15.2.dev0"
+current_version = "0.16.1.dev0"
 parse = '(?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\.(?P<release>[a-z]+)(?P<candidate>\d+))?'
 serialize = [
     '{major}.{minor}.{patch}.{release}{candidate}',

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/__init__.py

@@ -8,7 +8,7 @@ __author__ = 'DataCebo, Inc.'
 __copyright__ = 'Copyright (c) 2022 DataCebo, Inc.'
 __email__ = 'info@sdv.dev'
 __license__ = 'BSL-1.1'
-__version__ = '0.15.2.dev0'
+__version__ = '0.16.1.dev0'
 
 import logging
 

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/_benchmark/credentials_utils.py

@@ -13,5 +13,5 @@ def sdv_install_cmd(credentials):
 pip install sdv-installer
 
 python -c "from sdv_installer.installation.installer import install_packages; \\
-install_packages(username='{username}', license_key='{license_key}')"
+install_packages(username='{username}', license_key='{license_key}', version='0.53')"
 """)

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/benchmark.py

@@ -268,6 +268,9 @@ def _setup_output_destination_aws(
                     f's3://{bucket_name}/{synth_folder}/'
                     f'{final_synth_name}_synthetic_data.{synthetic_data_extension}'
                 ),
+                'quality_report': (
+                    f's3://{bucket_name}/{synth_folder}/{final_synth_name}_quality_report.pkl'
+                ),
                 'benchmark_result': (
                     f's3://{bucket_name}/{synth_folder}/{final_synth_name}_benchmark_result.csv'
                 ),
@@ -335,6 +338,7 @@ def _setup_output_destination(
                 'synthetic_data': str(
                     synth_folder / f'{final_synth_name}_synthetic_data.{synthetic_data_extension}'
                 ),
+                'quality_report': str(synth_folder / f'{final_synth_name}_quality_report.pkl'),
                 'benchmark_result': str(synth_folder / f'{final_synth_name}_benchmark_result.csv'),
                 'metainfo': str(top_folder / f'metainfo{suffix}.yaml'),
                 'results': str(top_folder / f'results{suffix}.csv'),
@@ -569,6 +573,8 @@ def _compute_scores(
     compute_privacy_score,
     modality,
     dataset_name,
+    synthesizer_path=None,
+    result_writer=None,
 ):
     metrics = metrics or []
     if modality == 'single_table':
@@ -633,6 +639,8 @@ def _compute_scores(
         quality_report.generate(real_data, synthetic_data, sdmetrics_metadata, verbose=False)
         output['quality_score_time'] = calculate_score_time(start)
         output['quality_score'] = quality_report.get_score()
+        if synthesizer_path is not None and result_writer is not None:
+            result_writer.write_report(quality_report, synthesizer_path['quality_report'])
 
     if compute_privacy_score:
         start = get_utc_now()
@@ -724,6 +732,8 @@ def _score(
                 compute_privacy_score,
                 modality,
                 dataset_name,
+                synthesizer_path,
+                result_writer,
             )
 
             output['timeout'] = False  # There was no timeout
@@ -1293,6 +1303,7 @@ def benchmark_single_table(
                     <synthesizer_name>/
                         synthesizer.pkl
                         synthetic_data.csv
+                        quality_report.pkl
         show_progress (bool):
             Whether to use tqdm to keep track of the progress. Defaults to ``False``.
 
@@ -1704,6 +1715,7 @@ def benchmark_multi_table(
                     <synthesizer_name>/
                         synthesizer.pkl
                         synthetic_data.csv
+                        quality_report.pkl
         show_progress (bool):
             Whether to use tqdm to keep track of the progress. Defaults to ``False``.
 

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/result_explorer/result_explorer.py

@@ -96,7 +96,7 @@ class ResultsExplorer:
         return self._handler.list()
 
     def _get_file_path(self, results_folder_name, dataset_name, synthesizer_name, file_type):
-        """Validate access to the synthesizer or synthetic data file."""
+        """Validate access to the synthesizer, synthetic data or quality report file."""
         end_filename = f'{synthesizer_name}'
         if file_type == 'synthetic_data':
             # Multi-table synthetic data is zipped (multiple CSVs), single table is CSV
@@ -106,6 +106,8 @@ class ResultsExplorer:
                 end_filename += '_synthetic_data.csv'
         elif file_type == 'synthesizer':
             end_filename += '.pkl'
+        elif file_type == 'quality_report':
+            end_filename += '_quality_report.pkl'
 
         date = '_'.join(results_folder_name.split('_')[-3:])
         path_parts = [results_folder_name, f'{dataset_name}_{date}', synthesizer_name]
@@ -154,6 +156,31 @@ class ResultsExplorer:
         )
         return self._handler.load_synthetic_data(file_path)
 
+    def load_quality_report(self, dataset_name, synthesizer_name, results_folder_name=None):
+        """Load the quality report for a given dataset and synthesizer.
+
+        Args:
+            dataset_name (str):
+                The name of the dataset.
+            synthesizer_name (str):
+                The name of the synthesizer.
+            results_folder_name (str, optional):
+                The name of the results folder to load from. If not provided,
+                the latest run will be used. Defaults to None.
+
+        Returns:
+            sdmetrics.reports.QualityReport:
+                The quality report computed during the benchmark.
+        """
+        results_folder_name = results_folder_name or self._get_latest_run()
+        file_path = self._get_file_path(
+            results_folder_name=results_folder_name,
+            dataset_name=dataset_name,
+            synthesizer_name=synthesizer_name,
+            file_type='quality_report',
+        )
+        return self._handler.load_quality_report(file_path)
+
     def load_real_data(self, dataset_name):
         """Load the real data for a given dataset.
 

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/result_explorer/result_handler.py

@@ -3,6 +3,7 @@
 import io
 import operator
 import os
+import tempfile
 import warnings
 from abc import ABC, abstractmethod
 from datetime import datetime
@@ -11,6 +12,7 @@ import cloudpickle
 import pandas as pd
 import yaml
 from botocore.exceptions import ClientError
+from sdmetrics.reports.base_report import BaseReport
 
 from sdgym._dataset_utils import _read_zipped_data
 from sdgym.benchmark import TIMEOUT, _add_adjusted_scores
@@ -58,6 +60,11 @@ class ResultsHandler(ABC):
         """Load synthetic data from a file."""
         pass
 
+    @abstractmethod
+    def load_quality_report(self, file_path):
+        """Load a quality report from a file."""
+        pass
+
     @abstractmethod
     def _load_yaml_file(self, folder_name, file_name):
         """Load a YAML file from the results folder."""
@@ -412,6 +419,10 @@ class LocalResultsHandler(ResultsHandler):
 
         return pd.read_csv(full_path)
 
+    def load_quality_report(self, file_path):
+        """Load a quality report from a pickle file."""
+        return BaseReport.load(os.path.join(self.base_path, file_path))
+
     def _get_results_files(self, results_folder_name, prefix, suffix):
         return [
             f
@@ -529,6 +540,17 @@ class S3ResultsHandler(ResultsHandler):
 
         return pd.read_csv(io.BytesIO(body))
 
+    def load_quality_report(self, file_path):
+        """Load a quality report from S3."""
+        key = f'{self.prefix}{file_path}'
+        response = self.s3_client.get_object(Bucket=self.bucket_name, Key=key)
+        with tempfile.TemporaryDirectory() as temp_dir:
+            temp_path = os.path.join(temp_dir, 'report.pkl')
+            with open(temp_path, 'wb') as temp_file:
+                temp_file.write(response['Body'].read())
+
+            return BaseReport.load(temp_path)
+
     def _get_results_files(self, results_folder_name, prefix, suffix):
         s3_prefix = f'{self.prefix}{results_folder_name}/'
         paginator = self.s3_client.get_paginator('list_objects_v2')

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym/result_writer.py

@@ -1,6 +1,7 @@
 """Results writer for SDGym benchmark."""
 
 import io
+import tempfile
 import zipfile
 from abc import ABC, abstractmethod
 from pathlib import Path
@@ -33,6 +34,11 @@ class ResultsWriter(ABC):
         """Write data to a YAML file."""
         pass
 
+    @abstractmethod
+    def write_report(self, report, file_path):
+        """Write an SDMetrics report to a pickle file."""
+        pass
+
 
 class LocalResultsWriter:
     """Local results writer for saving results to the local filesystem."""
@@ -95,6 +101,10 @@ class LocalResultsWriter:
         with open(file_path, 'wb') as f:
             cloudpickle.dump(obj, f)
 
+    def write_report(self, report, file_path):
+        """Write an SDMetrics report to a pickle file."""
+        report.save(file_path)
+
     def write_yaml(self, data, file_path, append=False):
         """Write data to a YAML file."""
         file_path = Path(file_path)
@@ -141,6 +151,14 @@ class S3ResultsWriter(ResultsWriter):
         buffer.seek(0)
         self.s3_client.put_object(Body=buffer.read(), Bucket=bucket, Key=key)
 
+    def write_report(self, report, file_path):
+        """Write an SDMetrics report to S3 as a pickle file."""
+        bucket, key = parse_s3_path(file_path)
+        with tempfile.TemporaryDirectory() as temp_dir:
+            temp_path = Path(temp_dir) / 'report.pkl'
+            report.save(temp_path)
+            self.s3_client.put_object(Body=temp_path.read_bytes(), Bucket=bucket, Key=key)
+
     def write_yaml(self, data, file_path, append=False):
         """Write data to a YAML file in S3."""
         bucket, key = parse_s3_path(file_path)

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0/sdgym.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdgym
-Version: 0.15.2.dev0
+Version: 0.16.1.dev0
 Summary: Benchmark tabular synthetic data generators using a variety of datasets
 Author-email: "DataCebo, Inc." <info@sdv.dev>
 License-Expression: BUSL-1.1
@@ -63,8 +63,8 @@ Requires-Dist: torch>=2.6.0; python_version >= "3.13" and python_version < "3.14
 Requires-Dist: torch>=2.9.0; python_version >= "3.14"
 Requires-Dist: tqdm>=4.66.3
 Requires-Dist: XlsxWriter>=1.2.8
-Requires-Dist: rdt>=1.18.2; python_version < "3.14"
-Requires-Dist: rdt>=1.20.0; python_version >= "3.14"
+Requires-Dist: rdt<2,>=1.18.2; python_version < "3.14"
+Requires-Dist: rdt<2,>=1.20.0; python_version >= "3.14"
 Requires-Dist: sdmetrics>=0.29.0
 Requires-Dist: sdv<2,>=1.37.0
 Provides-Extra: dask

{sdgym-0.15.2.dev0 → sdgym-0.16.1.dev0}/sdgym.egg-info/requires.txt

@@ -24,7 +24,7 @@ torch>=1.13.0
 
 [:python_version < "3.14"]
 cloudpickle>=2.1.0
-rdt>=1.18.2
+rdt<2,>=1.18.2
 
 [:python_version >= "3.10" and python_version < "3.11"]
 scikit-learn>=1.1.0
@@ -59,7 +59,7 @@ pandas<3,>=2.3.3
 scikit-learn>=1.8.0
 scipy>=1.16.1
 torch>=2.9.0
-rdt>=1.20.0
+rdt<2,>=1.20.0
 
 [all]
 sdgym[dask,dev,test]