Improve updating field_data_types in HyperTransformer

[npatki]

Expected behavior

• Rename field_data_types to sdtypes in the update methods
• In the update method, accept a dictionary with param name column_name_to_sdtype
• Always let the user know that the transformer will change based on the new updated sdtype

# updating the sdtype of a single column
>>> ht.update_sdtypes(column_name_to_sdtype={
       'colA': 'categorical'
 })
Info: The transformers for these columns may change based on the new sdtypes.
Use 'get_config()' to verify the transformers.

# updating the sdtypes of multiple columns
>>> ht.update_sdtypes(column_name_to_sdtype={
       'colA': 'categorical',
       'colB': 'categorical',
       'colC': 'categorical',
    })
Info: The transformers for these columns may change based on the new sdtypes.
Use 'get_config()' to verify the transformers.

Edge cases

Scenario 1: User tries to get or update the types without using auto_detect_config (see #399 )

>>> ht = HyperTransformer()
>>> ht.get_sdtypes()
{}
Tip: Use the `detect_initial_config` method to pre-populate all the sdtypes and transformers from your dataset.

# this will still work but will print out a tip
>>> ht.update_sdtypes(column_name_to_sdtype={
       'colA': 'categorical'
})
Tip: Use the `detect_initial_config` method to pre-populate all the sdtypes and transformers from your dataset.

Scenario 2: User updates the types after calling fit or fit_transform on the data already.

>>> ht.fit(data)
>>> ht.update_sdtypes(column_name_to_sdtype={
    'colA': 'numerical'
})
Info: The transformers for these columns may change based on the new sdtype.
Use 'get_config()' to verify the transformers.
Warning: For this change to take effect, please refit your data using 'fit' or 'fit_transform'.

Scenario 3: User tries to add in a type that is not available on open source

>>> ht.set_sdtypes(column_name_to_sdtype={
        'colA': 'credit_card'
})
Error: Unsupported sdtypes ('credit_card'). To use sdtypes with specific semantic meanings,
please contact the SDV team to update to rdt_plus. Otherwise, use 'pii' to anonymize the column.

[amontanez24]

For scenario 3, if some of the sdtypes provided are valid but one is invalid, should the valid ones be set? Or should the whole method crash?