Document the custom character removal heavily or provide a way to test the regexes

[bastienboutonnet]

When we roll out the objekt release the columns cleanups will be able to take a custom list of characters to remove by using regex string building.

If given a list like so:

characters_to_remove: ['a', '?']

we join then in Python by doing ''.join() this could be dangerous if people give \ for example as it will need to be escaped. So we need to make sure they know how to build it or that we don't concat it with \W type selectors for example.

This could get quite hairy so we might actually want to instead or generating the regex for them tell them to build a valid regex for them.

Also, it might be a good idea to look into how dbt people deal with it in their new release that is offering re somewhere in model selections and what not.