string_search_algorithms 1.1.0

Fast Dart library for string similarity and substring search with configurable engines, normalization, and caching. Pure Dart and null-safe.

Changelog

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

1.1.0

Added

• SimilarityAlgorithm.composite: a calibrated ensemble metric that combines Jaro-Winkler, Dice, Cosine, and Overlap witnesses into one stable score. Robust across typos, word reordering, and containment, and safe to use in findMatches / rankByRelevance because every pair is scored comparably.
• CompositeOptions and CompositeCombiner to configure the combiner (scaledMax, weightedMean, max), witness weights, and the per-witness calibration floors (defaults derived empirically; see benchmark/composite_calibration.dart).
• ExplainableMetric interface and ExplainedScore: metrics can now explain a score. compareWithDetails populates SimilarityResult.metadata with the composite's per-witness sub-scores, the combiner, and the dominant witness.

1.0.1

Documentation

• Refresh package description and README introduction copy.
• Update README install snippet for 1.0.1.

1.0.0

Added

• Initial stable release of the similarity and search libraries.
• Instance-based engines with configurable normalization and caching.
• Similarity algorithms: Dice, Levenshtein, Damerau-Levenshtein, OSA, Jaro, Jaro-Winkler, Cosine, Jaccard, Overlap, Tversky, N-gram, Hamming, LCS, Soundex, Metaphone.
• Substring search algorithms: KMP, Boyer-Moore, Rabin-Karp.
• CompiledPattern for repeated substring searches.
• Convenience String extension methods for similarity and search.
• Benchmark scripts under benchmark/.

Documentation

• Complete API documentation for all public members.
• Updated README with examples and configuration guidance.