2.0.1

What's Changed

• Clamp document_frequency in bm25 score before_each(). by @tmpfs in #25

Full Changelog: 2.0.0...2.0.1

2.0.0

What's Changed

• Update hashbrown dependency by @tmpfs in #23

Full Changelog: 2.0.0-alpha-2...2.0.0

Stable release

Release of 2.0.0 after a period of alpha testing

2.0.0-alpha-2

• Breaking: Field extractors are now multi-value typed by default. i.e, resolved value is a Vec of values rather than Some value

2.0.0-alpha-1

• Breaking change: add_document_to_index now accepts documents as references rather than values. The Tokenizer and Filter type now returns slices rather than owned strings

1.2.4

• Improved zero-to-one scoring
• Optimization in indexing

1.2.3

• Bugfix: Zero-To-One scoring, prevent scoring surpassing 1

1.1.3

• Performance improvement by representing index with arena instread of smart pointers

1.1.2

• Performance improved by using Mutex instead of RwLock for tree nodes
• Benchmarks

1.1.1

• Bug fix, empty tokens from tokenizer causing panic

1.1.0

• Thread safe with Arc/RwLock index