TermDictionaryFstBuilder

Namespace

Rowles.LeanLucene.Codecs.Fst

Assembly

Rowles.LeanLucene.dll

Serialises a sorted term list into a compact byte-keyed v2 dictionary format. Format: [termCount:int32][postingsOffsets: N*int64][keyStarts: (N+1)*int32][keyData: UTF-8 bytes]. This format enables O(log N) binary search on raw UTF-8 bytes without string materialisation. Terms are re-sorted in UTF-8 byte order to ensure binary search correctness (string ordinal sort can differ from UTF-8 byte sort for supplementary characters).

internal static class TermDictionaryFstBuilder

TermDictionaryFstBuilder

Methods

Write(IndexOutput, List<string>, Dictionary<string, long>)

Writes sorted terms and their postings offsets in v2 format to the given output. The codec header (magic + version) must already have been written by the caller.