Difference between data-import app and condition implementation

When doing record QA on our recent migration load, one of Stanford's staff members [noted](https://github.com/sul-dlss/FOLIO-Project-Stanford/issues/356) a difference between how the **mod-data-import** imports records using our Nolana mappings and how `folio_migration_tools` implements the same condition. For the `remove_prefix_by_indicator` condition for the `indexTitle`, the following 245 field:

```=245 10$a!Abracadabra! :$ba novel /$cby Wolf Mankowitz.```

is returned as '!Abracadabra! : a novel' without the trailing backslash due to the regular expression [substitution]
(https://github.com/FOLIO-FSE/folio_migration_tools/blob/64f8519afe62418fcc47857f30ae3c6bdc691a00/src/folio_migration_tools/marc_rules_transformation/conditions.py#L236-L238).

Using the data-import app, the trailing slash for the `indexTitle` is retained when importing the MARC record into FOLIO. What is the reasoning behind this logic? It seems to extend beyond the intention of this condition and if institutions wish to remove the trailing backslash, they could just add the `remove_ending_punc` condition to this rule.