Added table-based parser for is_alphanumeric/etc functions

[AndrewVallette]

Using a table-based parser provides better performance than the is_alpha/is_digit/etc functions in nom. This is due to the reduced number of comparisons that are performed for each character.

nom's functions work by checking if a character is within a specified range (or set of ranges), which requires multiple comparisons. The table-based lookup works by checking if the character is in one of the classes defined below, which requires only a single comparison. The table is very small (256 bytes) and is likely to remain in cache.

This can be expanded upon in the future (e.g., refactoring selector_parse to use a table), which will provide another noticeable increase in performance.