All notable changes to the Customized SymSpell project.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Ligature normalization - Automatic Unicode NFKC conversion (fi → fi)
- Case preservation - Maintains uppercase on first character in word segmentation
- Hyphen removal - Better handling of syllabification artifacts
- Punctuation adjacency - Punctuation stays adjacent to words
- Safety checks - Length validation in character operations
- Test suite - 19 new tests for v6.7 features (59 total, all passing)
- Enhanced
wordBreakSegmentation()method with v6.7 improvements - Improved accuracy for mixed-case input and special characters
- Maintained: 9,523 queries/second
- Accuracy: 88.61% (Qwerty keyboard distance)
- Overhead: < 1% from new features
All 5 custom features remain fully functional:
- Exclusion Dictionary
- Edit Factor Configuration
- Ignore Unknown Words
- Bigram Key Splitting
- QwertzDistance (German keyboard)
- Java port of SymSpell v6.6
- Weighted Damerau-Levenshtein distance
- Keyboard-based distance (Qwerty/Qwertz)
- Bigram dictionary support
- Compound word correction
- Word segmentation
- Custom exclusion dictionary
For detailed release information, see ReleaseNotes.md