Currently, ReprHighlighter handles numeric literals well, but it treats trailing measurement units (like GB, ms, %) as plain text. This makes technical logs and performance metrics harder to scan quickly, as the unit doesn't share the same visual importance or distinct styling as the number it belongs to.
I've developed an enhancement for the number regex in ReprHighlighter to include an optional number_unit named group. This allows Rich to recognize and style common units (TB, GB, MB, KB, B, ms, s, m, h, %) when they follow a numeric literal.
Proposed Regex:
r"(?P<number_unit>(?<!\w)\-?(?:0[xX][0-9a-fA-F_]+|0b[01_]+|(?:(?:\d{1,3}(?:[_,]\d{3})+)|\d+)(?:\.\d*)?(?:e[-+]?\d+)?)\s?(?:TB|GB|MB|KB|B|ms|h|s|m|%))"