Light normalization preprocessing for browse terms

[yossariano]

• As pointed out by Nora, our Elasticsearch folding analyzers don't seem to account for certain characters, hence we see things like this which look identical but have an invisible character difference:
  • Has results: https://www.nypl.org/research/research-catalog/browse/authors?q=Ki%C5%A1%2C%20Danilo%2C%201935-1989&search_scope=starts_with&sort=termLabel&sort_direction=asc
  • No results: https://www.nypl.org/research/research-catalog/browse/authors?q=Kis%CC%8C%2C%20Danilo%2C%201935-1989&search_scope=starts_with
• This change adds extremely light normalization into the preprocessing for the query which addresses the above discrepancy.