Module trigram 

Byte trigrams: the atomic unit of the candidate index.

A trigram is any three consecutive bytes of a file’s contents. A file is represented in the index by the set of distinct trigrams it contains; a query is represented by a boolean formula over trigrams that every matching file must satisfy. See docs/index-and-storage.md.

Functions

distinct

The set of distinct trigrams in bytes. A BTreeSet keeps results deterministic for tests; the indexer will use a faster set.

for_each

Invoke f once per (not-necessarily-distinct) trigram window over bytes.

of_literal

The trigrams of a literal byte string, in order (with repeats). Empty if lit.len() < 3.

pack

Pack a trigram into the low 24 bits of a u32, for compact keys.

Type Aliases

Trigram

A single trigram. Three raw bytes; works for arbitrary (incl. non-UTF-8) content.