Create a utility to convert old --> new metadata format

[npatki]

Problem Description

Since SDV 1.0 will use a slightly modified metadata format, it can be time-consuming to convert preexisting metadata files manually.

Expected behavior

Create utility methods that would convert metadata in the old format to valid metadata in the new format. Add a method called upgrade_metadata to each class that takes in the old metadata file and writes a new one.

from sdv.metadata import SingleTableMetadata

SingleTableMetadata.upgrade_metadata(
  old_filepath='old_metadata.json',
  new_filepath='new_metadata.json'
)

# same method would exist for MultiTableMetadata

Cases

Filepaths: The file in old_filepath must be found and there must be no file already existing in new_filepath.

SingleTableMetadata.upgrade_metadata(
  old_filepath='data/incorrect_path.json',
  new_filepath='new_metadata.json'
)
Error: No metadata file found at old filepath 'data/incorrect_path.json'

SingleTableMetadata.upgrade_metadata(
  old_filepath='old_metadata.json',
  new_filepath='old_metadata.json'
)
Error: A file already exists at path 'old_metadata.json'. Please specify a new filepath.

Metadata Validity. The user's old metadata may not be valid. After writing the new file, do the validate() check and if anything is found, throw a warning with the details.

from sdv.metadata import SingleTableMetadata

SingleTableMetadata.upgrade_metadata(
  old_filepath='old_metadata.json',
  new_filepath='new_metadata.json'
)

Warning: Successfully converted the old metadata, but the metadata was not valid.
To use this with the SDV, please fix the following errors.

InvalidMetadataError: The metadata is not valid

Error: Invalid values ("pii") for datetime column "start_date".
Error: Invalid regex format string "[A-{6}" for text column "user_id"
...

Single & Sequential Metadata. Some users may have written a single table or sequential metadata file using the multi-table format. As in, they have nested the (single) table under the "tables" keyword. We should properly convert the metadata in this case.

{
    "tables": {
        "my_table": { <table metadata> }
    }
}

However, if there are multiple table specified, then throw an error.

Error: There are multiple tables specified in the JSON.
Try using the MultiTableMetadata class to upgrade this file.

Conversion Error: Do not write the file if there are any others errors in parsing or converting the old metadata. Let whatever error occurs be raised.