I Wrote 82 Regex Replacements to Parse 6,933 Time Format Variations From a Government Dataset

A Japanese government pharmacy dataset has free-text business hours with 6,933 unique formats. This is the story of parsing them — and knowing when to stop.