Smaller Representation of Compiled Regular Expressions

초록

We consider the problem of running the regex pattern matching in a space-efficient manner. Given a regex, we suggest a bit-packing scheme for representing a compiled regex in a compressed way, which is its position automaton. Our scheme reduces its representation size further by relying on the homogeneous property of position automata and practical features of regexes. We implement the proposed scheme and evaluate the memory consumption using a practical regex benchmark dataset. Our approach produces a much smaller representation compared to two common FA representations. In addition, experimental results show that our bit-packing regex engine is effective for matching regexes that have large compiled forms, by showing less memory consumption compared to the current state-of-the-art regex engine (RE2).

키워드