novident_spell_checker 1.0.1

Spelling correction & fuzzy search based on the symmetric delete spelling correction algorithm. Dart port of symspell-ex with some extra things.

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

1.0.1 - 2026-08-15

• fix(tokenizer): preserve exact whitespace and token offsets by @CatHood0 in https://github.com/Novident/novident_spell_checker/pull/1

New Contributors

• @CatHood0 made their first contribution in https://github.com/Novident/novident_spell_checker/pull/1

1.0.0 - 2026-08-15

Initial release: Dart/Flutter port of symspell-ex (symmetric delete spelling correction) with a full dictionary ecosystem.

Added

Spell-check engine

• SymSpellEx: symmetric delete index with lookup/search (fuzzy suggestions), correct (full sentence correction with punctuation, spacing and capitalization preservation), add/train/trainDictionary (bulk training), hasTerm (O(1) exact membership) and clear.
• Synchronous API (the original is async only because of its Redis store); per-keystroke hot path is allocation-light.
• Multi-language dictionaries with isolated namespaces and runtime language switching.
• MemoryStore behind a small DataStore abstraction.
• DamerauLevenshteinDistance (transposition-aware, cached arrays).
• CoreTokenizer: url / number / word / punctuation classification for latin and arabic alphabets, with spacing metadata (token.distance).
• Trie: compact packed child/sibling vocabulary with prefix queries (withPrefix) for validity checks and autocomplete.
• Constants: ~190 Languages codes, TokenTags, Alphabets.

Dictionary formats & I/O

• Dictionary: plain word lists, word frequency lists and term,frequency CSV — parse (fromLines/fromCsv/fromBytes) and export (toLines/toBytes/toByteStream) with lossless round-trips.
• HunspellDictionary: .dic and personal dictionary parsing (count header, /FLAGS, escaped slashes, multi-word pairs, *forbidden entries, UTF-8 BOM) plus export back to Hunspell format.
• AffixRules: optional .aff support — FLAG (default/long/num/ UTF-8), PFX/SFX with stripping, conditions (., [abc], [^abc], multi-char), cross products and continuation classes. HunspellDictionary.expand(AffixRules) generates full word forms (measured: es_ES 55,820 stems → 652,256 forms, 11.7x).
• AssetDictionaryLoader: load, loadHunspell, loadAffix, loadHunspellWithAffixes for Flutter assets.
• Streaming I/O: fromByteStream/fromLineStream, trainByteStream/ trainLineStream, exportStream/toByteStream — chunk-boundary-safe reading and batched writing for very large dictionaries (1M+ words).

State management

• SymSpellEx.exportDictionary/exportLines/exportBytes/exportStream: export trained terms (including runtime-learned words) back to files; language-isolated and delete-key-free.