Feature/S3 to Postgres optional column order detection

[sleblanc23]

Description & motivation:

The s3_to_postgres and s3_dir_to_postgres callables currently rely on the S3 file(s) and the Postgres table to have matching column orders unless the order is specified in the column_customization config. This PR adds the option to set the column order by reading the header of the file in S3. It will be helpful for the upgrade of the load_heimdall_dag, currently the only user of this callable, to avoid errors due to column order.

I'm not convinced that I decided on the ideal implementation of this idea, with the user specifying the column delimiter to use the feature. I considered making the argument a boolean and parsing the delimiter from the options argument. That would be a simpler argument but would require some ugly regex, so I decided against it.

PR Merge Priority:

• Low

Tests and QC done:

• Created a test table in GSN with a different column order than the test file in S3 and confirmed that the table was loaded correctly
• Confirmed that excluding the new optional configuration does not affect exiting functionality