Practical Repetition-Aware Grammar Compression

Abstract: The goal of grammar compression is to construct a small sized context free grammar which uniquely generates the input text data. Among grammar compression methods, RePair is known for its good practical compression performance. MR-RePair was recently proposed as an improvement to RePair for constructing small-sized context free grammar for repetitive text data. However, a compact encoding scheme has not been discussed for MR-RePair. We propose a practical encoding method for MR-RePair and show its effectiveness through comparative experiments. Moreover, we extend MR-RePair to run-length context free grammar and design a novel variant for it called RL-MR-RePair. We experimentally demonstrate that a compression scheme consisting of RL-MR-RePair and the proposed encoding method show good performance on real repetitive datasets.