better handling of proper names in short def lookup

[balmas]

ᾍδης and other similar proper names that morpheus parses to lower case are annoying.

We get this weird loop:

lookup Αἵδης
no morpheus parse, form gives a see reference to ᾍδης
lookup ᾍδης
morpheus parses lemma as ᾅδης
get no definitions found because the entry ᾅδης in the short defs file points at the upper case entry for ᾍδης but our lemma is lower case so we get nothing.

Need to think about the best way to deal with this. Moving this to a new issue.

Originally posted by @balmas in alpheios-project/mjm#6 (comment)

[balmas]

fixing this one is going to be too much for the incremental.

We need better, and language-specific, fall-back handling altogether on the short definition lookups.

I was thinking we could perhaps just lookup the capitalized version of an entry when the lower case entry clearly pointed to a capitalized version, and if the target word was the same as the lemma, just capitalized.

But that doesn't work for Αἵδης, because of additional unicode complications.

morpheus reports ᾅδης as the lemma for ᾍδης . The short definition, following our standard approach, is indexed under the lowercase, and points at the upper case. This works when morpheus respects the upper case for the lemma, but when it doesn't we end up in a little bit of a circuitous loop. Further complicating things, in javascript, when you try to capitalize ᾅδης, it doesn't give you ᾍ as the capital of ᾅ, it follows the unicode entry which says it's Ἅ (U+1F0D) and the iota is added as a separate character. so now we can't even check to see if the targetWord as entered matches an uppercase version of the indexed lemma, because there is a mismatch on how it's upper-cased.