feat: Create parse-langtags utility to convert langtags.json

[darcywong00]

Fixes #14

The first part of this PR is creating a /DevTool/parse-langtags tool which converts the langtags.json file into a | separated language list that Speech Analyzer uses. (We won't revamp the language picker, so it will continue to use the iso639.txt file). The parse-langtags tool runs from node with

node dist/index.js -i iso639.txt

Note. The previous iso639.txt file had roughly 500 entries, and the langtags.json sourced file now has roughly 2000 entries. This causes a noticeable performance hit on "Exporting to Lift" as the iso639.txt file gets read in. I contemplated only populating the two columns we care about, but decided not to so users on the older versions of Speech Analyzer can just "drop in" the updated iso639.txt file.

---

Since langtags.json uses BCP-47 tags which can include script, region, and private use variants, the second part of this PR is updating Speech Analyzer code to accommodate this. (SA previously used 2-character country codes and language names). Sa_Doc.cpp now also has to merge how the "phonetic" (fonipa-x-etic) and "phonemic" (fonipa-x-emic) language tags get constructed.

• [x] There's still some encoding weirdness in the dropdown boxes. Even though the generated iso639.txt files are UTF-8, I suspect Speech Analyzer's usage of wstring means underlying data is in UCS-2 (not exactly UTF-16?)

Update - This is now displayed correctly with the rewrite of how iso639.txt gets parsed into codes.

Reference https://stackoverflow.com/questions/2527720/confused-about-cs-stdwstring-utf-16-utf-8-and-displaying-strings-in-a-win

[megahirt]

@darcywong00 did you intend to commit DistFiles/Fonts/DoulosSILR.ttf ?

[darcywong00]

At one point, I was trying to use an updated version of Doulos to render the dropdown box, but that didn't help. I'll revert that change

[darcywong00]

about x-bad-mru-Cyrl-RU and why we find it in langtags.json

He said it comes from CLDR, so langtags.json keeps it to stay in sync.