Support katakana stemming

[hackartisan]

Example from Martin H:

"Katakana: does アメリカ (amerika) finds アメリカン (amerikan)? It should, it doesn't, but practical consequences are not great (unlike hiragana, if two words in katakana follow each other, there usually is a middle dot in between, which is recognized as a word boundary.)"

We could add bigrams for this use case, or use specialized parsing from Stanford: https://github.com/sul-dlss/sul-solr-configs/blob/master/searchworks-prod/schema.xml#L493