Fun fact of the day:

In Germany, A Big Mac used to be a Big Mäc.

And the Filet-o-Fish was the Fishmäc. The spelling with the umlaut actually gets German speakers a little closer to the English pronunciation of “Mac.” But in 2007 McDonald’s took away the umlauts, and now Germans have to get boring ol’ Big Macs like the rest of us.

This PR would break some existing behavior, eg:

records indexed as Österreich would be query-able using the input Oesterreich but not Osterreich.

prior to this PR the term Österreich was retrievable using the term Osterreich so this would be a breaking change.

It seems like a better strategy would be to import two versions of the same token; once as Oesterreich and once as Osterreich.

Do we want to close this PR in favor of something icu related?

I don't believe the ICU library will handle this type of expansion.

This feature will require alt-names as the country name above can have 3 forms:

Österreich
Oesterreich
Osterreich

moving to alt-names feature as it can only be solved for a maximum of 2 cases before then.

Closed due to teh new hotness