The decimal is split into whole and fractional parts, then the fractional part is converted using a fixed precision and reduced with the greatest common divisor.

This makes the result simple to read while still rejecting empty or malformed input.