Hyperglot

- FIX: Improved character set parsing from database properly decomposing any combining characters into their parts and checking against those
- TESTS: Added first pytest for above case

- FEATURE: Added `--strict-support` flag (default False) to explicitly trigger warning about languages with unconfirmed status. Since those languages have well researched charset information but just have not been confirmed by several expert sources we still want to include them in the count. Using `--strict-support` excludes (but lists separately) all those languages which we have not been able to confirm
- TWEAK: Renamed `--strict` flag to `--strict-iso` to be more discriptive
- TWEAK: Database file linking, one more time... as per 0.1.10
- TWEAK: Added validation check to prevent non-space separators in character list data

- FEATURE: Implemented `fontlang-export` CLI script to export the rosetta.yaml with expanded inherits to a file, usage: `$ fontlang-export thefile.yaml`

- FIX: Refactored `setup.py` to include the databased file relative to the package

- FIX: "Inverted" the `preferred_as_individual` outcome, e.g. those languages should suppress any included languages from being listed and be listed as one language instead

- FIX: Made sure `preferred_as_individual` in fact also removes the language that is being inheritted from the matches
- TWEAK: Update `fontlang-save` and sorting to not include inheritted attributes
- TWEAK: Updated and fixed validation for `status` attributes