Hyperglot

- DATA: Added Benin languages
- DATA: Small fix to Portuguese

- DATA: Revised Tamil orthography
- DATA: Added Apinayé, Karo and Awetí languages

- FIX: Fixed an encoding issue affecting Windows environments
- DATA: Fixed typos in Buginese
- DATA: Reviewed Minangkabau orthography

- DATA: Added Batak languages and refined Balinese
- FIX: Further improvement to detection of orthographies with unencoded base + mark combinations
- TWEAK: Refined the returned properties of `hyperglot.language.Orthography` to include base and auxiliary lists of _encoded_ characters as well as _required_ marks for
- TOOLS: Added scraper for fetching a mapping of Opentype language systems to ISO codes and saving them in `other/languagesystems.yaml`

- DATA: Renamed `design_note` to `design_requirements` and made its data structure a list
- DATA: Introduced `design_alternates` - a list of characters which may require special design in a font supporting an orthography
- DATA: Added `design_alternates` for several Cyrillic and Latin languages

- DATA: Corrected speaker count for Manipuri
- DATA: Updates to Andaandi and Old Nubian
- DATA: Minor formatting and duplicate fixes
- FIX: Fixed parsing issue that led for some languages to require marks in their support _as if_ the `--marks` flag was used
- TWEAK: `hyperglot.language.Language` no longer prunes or parses any character lists, but this is instead done on running the support checks by instantiating a `Orthography` object and using it for checking, leaving the dict representation of the yaml data in the `Language` untouched
- FEATURE: Introduced `hyperglot.language.Orthography` abstraction for easier access of parse lists vs yaml raw character strings
- TESTS: More refactored Languages, Language and new Orthography tests