* Added Hindi language
* Removed the hardcoded special characters from language validation. Now reading them from the .yml
* improved method of hiding the letters on 0 and 1, when needed
* virtual keypad adjustments
* improved the single-letter validation during build time
* improved Devanagari validation script
* improved sorting when filters are on
* added Korean language
* fokin context no more messing up everything in the InputModes
* no more unnecessary textField and inputType passing in the InputModes
* a single source of truth for the InputMode kind
* ModePredictive -> ModeWords
* no more db queries to increase the priority of emojis and special chars
* Korean virtual keypad
* more consistent displaying of the ABC string
* sorted out the labels of 1-key and 0-key in numeric modes
* documentation update
* new dictionary format that supports syllabaries
* optimized the dictionary build cache significantly to truly build only the changed language files
* code style fixes
