Labeled matches: why is this not in every regex engine?
Summary
The article explores labeling and named entity recognition using regex patterns and DFA-based automata (RE#), comparing performance against spaCy's NER. It provides real-world examples (dates, money, emails, URLs) and demonstrates how large regexes can be compiled into efficient state machines, including open-source code in the RESHARP project. It also discusses advanced regex features like intersection and complement to improve accuracy and speed.