5.0.0

What's Changed

• Much of the tokenizer rewritten for speed (its at least 3x as fast)
• Move code previously using xlsx option into a version with a separate entry point: @borgar/fx/xlsx
• Split up translateToA1 into string and token versions: translateFormulaToA1 & translateTokensToA1
• Split up translateToR1C1 into string and token versions: translateFormulaToR1C1 & translateTokensToR1C1
• Translation functions made xlsx agnostic
• Expose a fast function for range specific parsing: parseR1C1Range & parseA1Range
• Library moved to TypeScript and types cleaned up as possible
• Docs are now built using typedoc which make them play better with githubs hashes