Version 3.4.3

• Fixed #261
• Rewrote tests to Jest framework
• Wrote tests for TypeScript typings
• Cleanup build

Version 3.4.2

• Fixed #288

Version 3.4.1

• Ensured dist/ content is production ready (both full and min versions) #283

Version 3.4.0

• Upgraded build tool to Webpack 4. New dist/ output.

Version 3.3.1

• Fixed the circular JSON TypeError (#197). Thanks ThinkTankShark!

Version 3.2.1

• Fixed issue in which more fuzzy matches would weaken a score instead of strengthening it (#233)

Version 3.2.0

• Give better result for exact match when using weighted keys (#192)

Version 3.1.0

• Added match index location for array key (#183)
• Allow searching deep nested numbers (#189)

Version 3.0.5

• Escape special characters in search pattern (#168)

Version 3.0.4

• Random bug fixes (#162)

Version 3.0.0

• Removed Bower support
• Modified library into a more more palatable architecture, where the Bitap portion is now its own separate module.

BREAKING CHANGES

• Removed include option in favor of more explicit booleans: includeScore and includeMatches. Both are false by default.
• Removed searchFn option, as this (for now) will remain a Bitap based solution

Version 2.7.4

• Reverted to previous version, thus fixing breaking changes (a little bit of a version match here)

Version 2.6.2

• Revert back to previous version

Version 2.6.2

• Fix typings based on TypeScript guidelines (#129)

Version 2.6.0

• Added Typescript definition
• Added ability to set min/max matched character lengths when returning the matched indices (#122)

Version 2.5.0

• Added option to search by matching all tokens (in every record) when matchAllTokens:true (#95)

Version 2.3.0

• Added token separator to options, when tokenize:true (#93)
• General code clean up (#88)
• Bunch of other bug fixes

Version 2.2.0

• Added option to include matched indices (#6)
• Added ability to search with weighted keys (#62)

Version 2.1.0-beta

• Added ability to search with weighted keys (#62)

Version 2.0.0

• Modified search algorithm to search individual words AND the full string, computing the final score as a function of both. This yields better scoring accuracy (#41)
• Changed exact substrings to not have a score of zero. That is searching for "hell" in "hello" will not yield a score of zero, while searching for "hello" will (#63)
• Added verbose option, which will print to the console useful information, mostly for debugging
• Improved code structure.
• Added version information within Fuse itself
• Added this Changelog (#64)
• Added fallback when pattern length is greater than machine word length (i.e, > 32 characters) (#38)
• Allowed results with a value of 0 to be returned (#73)