Make synthesizers work with column_relationships

[amontanez24]

Problem Description

As a user, I expect my synthetic data to appropriately maintain column relationships if I specify it in my metadata.

Expected behavior

Synthesizer Initialization

• Raise a warning if column relationships are provided in the metadata but the user doesn't have permissions to use the advanced transformers

synthesizer = HMASynthesizer(metadata)

Warning: The metadata contains a column relationship of type 'address'. This relationship will be ignored. For higher quality data in this relationship, please inquire about the SDV Enterprise tier.

Transformer assignment

• For the address type, we should automatically assign the RandomLocationGenerator to the set of columns. However, the user would be able to update this using update_transformers – just like they would with any other transformer.

>>> synthesizer.auto_assign_transformers(data)
>>> synthesizer.get_transformers()
{
  'age': FloatFormatter(...),
  'credit_card': UniformEncoder(),
  ('vendor_city', 'vendor_state', ...): address.RandomLocationGenerator()
}

>>> from rdt.transformers.address import RegionalAnonymizer
>>> synthesizer.update_transformers({
  ('vendor_city', 'vendor_state', ...): RegionalAnonymizer()
})

• Deprecate the set_address_column method
  • No need to do any logic in this method. Just raise the warning.

>>> metadata = MultiTableMetadata.load_from_json(...)
>>> synthesizer = HSASynthesizer(metadata)
>>> synthesizer.set_address_columns(...)
Warning: set_address_columns is deprecated. Please add these columns directly to your metadata using 'add_column_relationship'

Additional context

• This is blocked by Add add_column_relationship method to single table metadata #1699
• Make sure that constraints error if any column is in a column relationship