Security fixes and minor updates

• Security fix for cve vulnerabilities
• Minor fixes for error logging and slf4j logging
• improved test coverage for custom tokenizers

Java 11 support

• Support for Java 11
• Significant improvement in memory footprint

Support to Chain multiple tokens

• Support for chaining multiple Tokenizers using chainTokenizers function #54

Date Range Fixes

• DateMatch with NeighborhoodRange greater than 0.9 #35
• Security fixes and updates on dependent libraries

Support for AGE ElementType

• Support for AGE ElementType
• Security fixes and updates on dependent libraries

Support for Grouping Matches

• New service method to return groups of Matches E.g. A matches with B and B with C, all of them appear in the same group (set of matches)
• Generic support for Tokens
• Enhanced support for jnbridge proxies
• NEAREST_NEIGHBORS match type miscellaneous fixes

1.0 Release

• Significant Performance increase from 0.x version. Provides a linear cost for most types of data
• From earlier version we can see at least 4X improvements in response times for match
• Reduces memory usage to more than 1/2 from earlier version
• Replaces configuration of SimilarityMatchFunction and MatchOptimizerFunction in Element to a more simplified MatchType and provides performance guarantees
• NUMBER and DATE types are now matched using Nearest Neighbor algorithm
• Element is now a Generic, to allow support for wider range of Data Types in the future

Memory Optimization

• Improved memory utilization by processing all elements by it's type or classification. This allows jvm to run gc to free heap space after each element type is processed
• Ability to configure scoring function in Element #19

Reducing memory usage

• ScoringFunction changed to BiFunction from Function. It takes a List of childScores as the second input. This allows childScores to be stored in function stack when score is evaluated instead of the heap, and reduces the overall memory footprint of the library
• Element.Builder.setValue overridden to accept Object Type
• Match.childScores is dropped to reduce memory footprint.

Memory management

Reduced memory footprint by removing tokens not meeting match threshold