Try it online at sherlocode.com !

This is a straightforward implementation of Regular Expression Matching with a Trigram Index by Russ Cox. For each of the characters, bigrams and trigrams, we store the set of line numbers where they appear. To answer a query, we first approximate the regex by enumerating the ngrams that are required/optional: the intersection/union of their indexed set yields good line candidates against which we can match the query.