sdv 1.36.4.dev0__tar.gz → 1.37.1.dev0__tar.gz

{sdv-1.36.4.dev0/sdv.egg-info → sdv-1.37.1.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdv
-Version: 1.36.4.dev0
+Version: 1.37.1.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-1.36.4.dev0 → sdv-1.37.1.dev0}/pyproject.toml

@@ -149,7 +149,7 @@ namespaces = false
 version = {attr = 'sdv.__version__'}
 
 [tool.bumpversion]
-current_version = "1.36.4.dev0"
+current_version = "1.37.1.dev0"
 parse = '(?P<major>\d+)\.(?P<minor>\d+)\.(?P<patch>\d+)(\.(?P<release>[a-z]+)(?P<candidate>\d+))?'
 serialize = [
     '{major}.{minor}.{patch}.{release}{candidate}',

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/__init__.py

@@ -6,7 +6,7 @@
 
 __author__ = 'DataCebo, Inc.'
 __email__ = 'info@sdv.dev'
-__version__ = '1.36.4.dev0'
+__version__ = '1.37.1.dev0'
 
 
 import sys

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/_utils.py

@@ -507,6 +507,16 @@ def warn_load_deprecated():
     )
 
 
+def warn_set_constraints_deprecated():
+    """Warn that the `set_constraints` method on synthesizer is deprecated."""
+    deprecation_msg = (
+        'Warning: The `set_constraints` method is deprecated. '
+        'Please use the `load_constraints` utility function to load constraints from a file '
+        'and add them to the synthesizer with the `add_constraints` method.'
+    )
+    warnings.warn(deprecation_msg, FutureWarning)
+
+
 def _validate_correct_synthesizer_loading(synthesizer, cls):
     """Validate that the loaded synthesizer is of the correct type."""
     synthesizer_name = synthesizer.__class__.__name__

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/cag/_utils.py

@@ -1,5 +1,7 @@
 import importlib
+import json
 import re
+import traceback
 import warnings
 
 import numpy as np
@@ -269,3 +271,29 @@ def load_constraint_from_dict(constraint_dict):
         raise ValueError(f"Unknown `constraint_class` '{class_name}'.")
 
     return constraint_class.load_constraint_from_dict(parameters=parameters)
+
+
+def _load_constraints_from_file(filepath):
+    """Load constraints from a file (JSON).
+
+    Args:
+        filepath (str):
+            The string path to the file containing the constraints to load.
+
+    Returns:
+        list[BaseConstraint]:
+            A list of constraint objects.
+    """
+    with open(filepath, 'r') as f:
+        constraints_json = json.load(f)
+
+    constraint_list = []
+    for constraint_dict in constraints_json:
+        try:
+            constraint_list.append(load_constraint_from_dict(constraint_dict))
+        except Exception as e:
+            warnings.warn(
+                f'Could not load constraint ({constraint_dict}):\n'
+                f'    {traceback.format_exception_only(type(e), e)[0]}'
+            )
+    return constraint_list

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/datasets/demo.py

@@ -73,6 +73,27 @@ def _get_data_from_bucket(object_key, bucket, client):
     return response['Body'].read()
 
 
+def _save_data_from_bucket(object_key, bucket, client, output_filepath):
+    """Save a file from an S3 bucket to the output_file.
+
+    Args:
+        object_key (str):
+            The key of the object to get.
+        bucket (str):
+            The name of the bucket to get the object from.
+        client (botocore.client.S3):
+            S3 client.
+        output_filepath (str):
+            The location to save the file to disk.
+    """
+    parent = os.path.dirname(str(output_filepath))
+    if parent:
+        os.makedirs(parent, exist_ok=True)
+
+    response = client.download_file(Bucket=bucket, Key=object_key, Filename=output_filepath)
+    return response['Body'].read()
+
+
 def _get_dataset_name_from_prefix(prefix):
     return prefix.split('/')[1]
 
@@ -211,7 +232,7 @@ def _get_first_v1_metadata_bytes(contents, dataset_prefix, bucket, client):
     )
 
 
-def _download_text_file_error_message(
+def _download_file_error_message(
     modality,
     dataset_name,
     output_filepath=None,
@@ -666,7 +687,7 @@ def _save_document(text, output_filepath, filename, dataset_name):
         LOGGER.info(f'Error saving {filename} for dataset {dataset_name}.')
 
 
-@handle_aws_client_errors(_download_text_file_error_message)
+@handle_aws_client_errors(_download_file_error_message)
 def _get_text_file_content(
     modality, dataset_name, filename, output_filepath=None, bucket=PUBLIC_BUCKET, credentials=None
 ):
@@ -717,7 +738,53 @@ def _get_text_file_content(
     return text
 
 
-@handle_aws_client_errors(_download_text_file_error_message)
+@handle_aws_client_errors(_download_file_error_message)
+def _save_file_content(
+    modality, dataset_name, filename, output_filepath, bucket=PUBLIC_BUCKET, credentials=None
+):
+    """Fetch text file content under the dataset prefix.
+
+    Args:
+        modality (str):
+            The modality of the dataset: ``'single_table'``, ``'multi_table'``, ``'sequential'``.
+        dataset_name (str):
+            The name of the dataset.
+        filename (str):
+            The filename to fetch (``'README.txt'`` or ``'SOURCE.txt'``).
+        output_filepath (str):
+            Save the file contents at this path.
+        bucket (str):
+            The name of the bucket to download from. Only 'sdv-datasets-public' is supported in
+            SDV Community. SDV Enterprise is required for other buckets.
+        credentials (dict):
+            Dictionary containing DataCebo license key and username. It takes the form:
+            {
+                'username': 'example@datacebo.com',
+                'license_key': '<MY_LICENSE_KEY>'
+            }
+    """
+    _validate_modalities(modality)
+    if os.path.exists(str(output_filepath)):
+        raise ValueError(
+            f"A file named '{output_filepath}' already exists. Please specify a different filepath."
+        )
+
+    dataset_prefix = f'{modality}/{dataset_name}/'
+    s3_client = _create_s3_client(bucket=bucket, credentials=credentials)
+    contents = _list_objects(dataset_prefix, bucket=bucket, client=s3_client)
+    key = _find_text_key(contents, dataset_prefix, filename)
+    if not key:
+        _raise_warnings(filename, output_filepath)
+        return None
+
+    try:
+        _save_data_from_bucket(
+            key, bucket=bucket, client=s3_client, output_filepath=output_filepath
+        )
+    except Exception:
+        LOGGER.info(f'Error saving {filename} for dataset {dataset_name}.')
+
+
 def get_source(
     modality, dataset_name, output_filepath=None, s3_bucket_name=PUBLIC_BUCKET, credentials=None
 ):
@@ -788,3 +855,44 @@ def get_readme(
         bucket=s3_bucket_name,
         credentials=credentials,
     )
+
+
+def save_resource(
+    modality,
+    dataset_name,
+    resource_filename,
+    output_filepath,
+    s3_bucket_name='sdv-datasets-public',
+    credentials=None,
+):
+    """Save the resource to disk.
+
+    Args:
+        modality (str):
+            The modality of the dataset: ``'single_table'``, ``'multi_table'``, ``'sequential'``.
+        dataset_name (str):
+            The name of the dataset to get the README for.
+        resource_filename (str):
+            The name of the file to download from S3.
+        output_filepath (str or None):
+            Optional path where to save the file.
+        s3_bucket_name (str, optional):
+            The name of the bucket to download from. Only 'sdv-datasets-public' is supported in
+            SDV Community. SDV Enterprise is required for other buckets. Defaults to
+            'sdv-datasets-public'.
+        credentials (dict, optional):
+            Dictionary containing DataCebo license key and username. It takes the form:
+            {
+                'username': 'example@datacebo.com',
+                'license_key': '<MY_LICENSE_KEY>'
+            }
+            Defaults to None.
+    """
+    _save_file_content(
+        modality=modality,
+        dataset_name=dataset_name,
+        filename=resource_filename,
+        output_filepath=output_filepath,
+        bucket=s3_bucket_name,
+        credentials=credentials,
+    )

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/multi_table/base.py

@@ -22,13 +22,14 @@ from sdv._utils import (
     check_synthesizer_version,
     generate_synthesizer_id,
     warn_load_deprecated,
+    warn_set_constraints_deprecated,
 )
 from sdv.cag._errors import ConstraintNotMetError
 from sdv.cag._utils import (
     _convert_to_snake_case,
     _get_invalid_rows,
+    _load_constraints_from_file,
     _validate_constraints,
-    load_constraint_from_dict,
 )
 from sdv.cag.programmable_constraint import ProgrammableConstraint, ProgrammableConstraintHarness
 from sdv.errors import (
@@ -328,18 +329,8 @@ class BaseMultiTableSynthesizer:
                 'Cannot `set_constraints` since constraints have already been applied.'
             )
 
-        with open(filepath, 'r') as f:
-            constraints_json = json.load(f)
-
-        constraint_list = []
-        for constraint_dict in constraints_json:
-            try:
-                constraint_list.append(load_constraint_from_dict(constraint_dict))
-            except Exception as e:
-                warnings.warn(
-                    f'Could not load constraint ({constraint_dict}):\n'
-                    f'    {traceback.format_exception_only(type(e), e)[0]}'
-                )
+        warn_set_constraints_deprecated()
+        constraint_list = _load_constraints_from_file(filepath)
 
         for constraint in constraint_list:
             try:

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/single_table/base.py

@@ -32,13 +32,14 @@ from sdv._utils import (
     check_synthesizer_version,
     generate_synthesizer_id,
     warn_load_deprecated,
+    warn_set_constraints_deprecated,
 )
 from sdv.cag._errors import ConstraintNotMetError
 from sdv.cag._utils import (
     _convert_to_snake_case,
     _get_invalid_rows,
+    _load_constraints_from_file,
     _validate_constraints_single_table,
-    load_constraint_from_dict,
 )
 from sdv.cag.programmable_constraint import ProgrammableConstraint, ProgrammableConstraintHarness
 from sdv.data_processing.data_processor import DataProcessor
@@ -552,18 +553,8 @@ class BaseSynthesizer:
                 'Cannot `set_constraints` since constraints have already been applied.'
             )
 
-        with open(filepath, 'r') as f:
-            constraints_json = json.load(f)
-
-        constraint_list = []
-        for constraint_dict in constraints_json:
-            try:
-                constraint_list.append(load_constraint_from_dict(constraint_dict))
-            except Exception as e:
-                warnings.warn(
-                    f'Could not load constraint ({constraint_dict}):\n'
-                    f'    {traceback.format_exception_only(type(e), e)[0]}'
-                )
+        warn_set_constraints_deprecated()
+        constraint_list = _load_constraints_from_file(filepath)
 
         for constraint in constraint_list:
             try:

sdv-1.37.1.dev0/sdv/utils/__init__.py

@@ -0,0 +1,15 @@
+"""Utils module."""
+
+from sdv.utils.utils import (
+    drop_unknown_references,
+    get_random_sequence_subset,
+    load_synthesizer,
+    load_constraints,
+)
+
+__all__ = (
+    'drop_unknown_references',
+    'get_random_sequence_subset',
+    'load_synthesizer',
+    'load_constraints',
+)

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0}/sdv/utils/utils.py

@@ -15,6 +15,7 @@ from sdv._utils import (
     check_synthesizer_version,
     generate_synthesizer_id,
 )
+from sdv.cag._utils import _load_constraints_from_file
 from sdv.errors import InvalidDataError, SamplingError, SynthesizerInputError
 from sdv.logging import get_sdv_logger
 from sdv.metadata.metadata import Metadata
@@ -198,3 +199,17 @@ def load_synthesizer(filepath):
     })
 
     return synthesizer
+
+
+def load_constraints(filepath):
+    """Load constraints from a file (JSON).
+
+    Args:
+        filepath (str):
+            The string path to the file containing the constraints to load.
+
+    Returns:
+        list[BaseConstraint]:
+            A list of constraint objects.
+    """
+    return _load_constraints_from_file(filepath)

{sdv-1.36.4.dev0 → sdv-1.37.1.dev0/sdv.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sdv
-Version: 1.36.4.dev0
+Version: 1.37.1.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-1.36.4.dev0/sdv/utils/__init__.py

@@ -1,5 +0,0 @@
-"""Utils module."""
-
-from sdv.utils.utils import drop_unknown_references, get_random_sequence_subset, load_synthesizer
-
-__all__ = ('drop_unknown_references', 'get_random_sequence_subset', 'load_synthesizer')