Remove the need to pass in Metadata to DataProcessor

sdv/data_processing/data_processor.py

@@ -24,7 +24,6 @@
 from sdv.data_processing.numerical_formatter import NumericalFormatter
 from sdv.data_processing.utils import load_module_from_path
 from sdv.errors import SynthesizerInputError, log_exc_stacktrace
-from sdv.metadata.metadata import Metadata
 from sdv.metadata.single_table import SingleTableMetadata
 
 LOGGER = logging.getLogger(__name__)
@@ -115,8 +114,6 @@ def __init__(
         locales=['en_US'],
     ):
         self.metadata = metadata
-        if isinstance(metadata, Metadata):
-            self.metadata = metadata._convert_to_single_table()
         self._enforce_rounding = enforce_rounding
         self._enforce_min_max_values = enforce_min_max_values
         self._model_kwargs = model_kwargs or {}

sdv/datasets/demo.py

@@ -4,6 +4,7 @@
 import json
 import logging
 import os
+import warnings
 from collections import defaultdict
 from pathlib import Path
 from zipfile import ZipFile
@@ -15,8 +16,6 @@
 from botocore.client import Config
 from botocore.exceptions import ClientError
 
-from sdv.metadata.multi_table import MultiTableMetadata
-from sdv.metadata.single_table import SingleTableMetadata
 from sdv.metadata.metadata import Metadata
 
 LOGGER = logging.getLogger(__name__)
@@ -114,6 +113,7 @@ def _get_metadata(output_folder_name, in_memory_directory, dataset_name):
     else:
         metadata_path = 'metadata_v2.json'
         if metadata_path not in in_memory_directory:
+            warnings.warn(f'Metadata for {dataset_name} is missing updated version v2.')
             metadata_path = 'metadata_v1.json'
         metadict = json.loads(in_memory_directory[metadata_path])
         metadata = metadata.load_from_dict(metadict, dataset_name)

sdv/metadata/metadata.py

@@ -1,7 +1,5 @@
 """Metadata."""
 
-from pathlib import Path
-
 from sdv.metadata.errors import InvalidMetadataError
 from sdv.metadata.multi_table import MultiTableMetadata
 from sdv.metadata.single_table import SingleTableMetadata

tests/integration/data_processing/test_data_processor.py

@@ -55,7 +55,7 @@ def test_with_anonymized_columns(self):
         metadata.update_column('adult', 'occupation', sdtype='job', pii=True)
 
         # Instance ``DataProcessor``
-        dp = DataProcessor(metadata)
+        dp = DataProcessor(metadata._convert_to_single_table())
 
         # Fit
         dp.fit(data)
@@ -111,7 +111,7 @@ def test_with_anonymized_columns_and_primary_key(self):
         data['id'] = np.arange(0, size).astype('O')
 
         # Instance ``DataProcessor``
-        dp = DataProcessor(metadata)
+        dp = DataProcessor(metadata._convert_to_single_table())
 
         # Fit
         dp.fit(data)
@@ -247,7 +247,7 @@ def test_prepare_for_fitting(self):
         data, metadata = download_demo(
             modality='single_table', dataset_name='student_placements_pii'
         )
-        dp = DataProcessor(metadata)
+        dp = DataProcessor(metadata._convert_to_single_table())
 
         # Run
         dp.prepare_for_fitting(data)
@@ -288,7 +288,7 @@ def test_reverse_transform_with_formatters(self):
         """End to end test using formatters."""
         # Setup
         data, metadata = download_demo(modality='single_table', dataset_name='student_placements')
-        dp = DataProcessor(metadata)
+        dp = DataProcessor(metadata._convert_to_single_table())
 
         # Run
         dp.fit(data)
@@ -327,7 +327,7 @@ def test_refit_hypertransformer(self):
         """Test data processor re-fits _hyper_transformer."""
         # Setup
         data, metadata = download_demo(modality='single_table', dataset_name='student_placements')
-        dp = DataProcessor(metadata)
+        dp = DataProcessor(metadata._convert_to_single_table())
 
         # Run
         dp.fit(data)
@@ -348,7 +348,7 @@ def test_localized_anonymized_columns(self):
         data, metadata = download_demo('single_table', 'adult')
         metadata.update_column('adult', 'occupation', sdtype='job', pii=True)
 
-        dp = DataProcessor(metadata, locales=['en_CA', 'fr_CA'])
+        dp = DataProcessor(metadata._convert_to_single_table(), locales=['en_CA', 'fr_CA'])
 
         # Run
         dp.fit(data)