16.2.0

• ENH: Add ability to set the the number of top/bottom values tracked (Issue #127)

16.1.0

• ENH: Add ability to set the maximum number of shapes tracked (Issue #127)
• INT: Bump guava to 33.4.5-jre

16.0.7

• ENH: Add new Semantic Type - IDENTITY.SIN_CA (Canadian - Social Insurance Number)
• ENH: Support --faker with inline specification (also fix incorrect example in README)
• INT: Improve error message if calling Distributions entry points with DISTRIBUTIONS not enabled
• INT: Add some more test cases for totalCount and merging
• INT: Bump google phonenumber to 9.0.1, logback-classic to 1.5.18

16.0.6

• ENH: Change the default number of Histogram bins to 200 from 1000 (improves performance by ~3x for large inputs), to revert to prior behavior invoked setHistogramBins(1000) on the AnalysisConfig. Note: If distributions (Quantiles/Histograms) are not required they can be disabled via configure() on the AnalysisConfig which will dramatically improve performance further.
• BUG: Honor minLength in FakerStringLT

16.0.5

• ENH: Add new Semantic Type - CRYPTOCURRENCY.TEXT_EN
• ENH: Faker - add String support to faker
• INT: Bump jackson to 2.18.3

16.0.4

• ENH: Add new Semantic Type - FILENAME_EXT
• ENH: ~15% improvement in throughput if not calculating distributions (~3% if calculating distributions)
• ENH: Change the default number of Histogram bins to 200 from 1000 (improves performance by ~3x for large inputs), to revert to prior behavior invoked setHistogramBins(1000) on the AnalysisConfig. Note: If distributions (Quantiles/Histograms) are not required they can be disabled via configure() on the AnalysisConfig which will dramatically improve performance further..
• BUG: Fix flaky test case - extendedTotalsDates
• INT: Bump google phonenumber to 9.0.0

16.0.3

• ENH: Faker - Switch to JSON definition, add String support to faker
• ENH: Improve the list of file extensions
• ENH: Minor performance improvements
• BUG: Fixed trivial bug in MergeSimple example
• BUG: Setting debug on the RecordAnalyzer template was not setting debug on the individual Analyzers
• BUG: Minor change to support building on Java 17
• ENH: Add memory diagnostics (debug level >= 3)
• INT: Bump gradle to 8.13, logback-classic to 1.5.17

16.0.2

• BUG: Most total* (except totalCount) not surviving merge (and associated tests)
• INT: Upgrade all examples to point to 16.+
• INT: Bump org.springframework.boot to 3.4.3

16.0.0

• ENH: Changed name of Feature NULL_AS_TEXT to NULL_TEXT_AS_NULL
• ENH: Added mini-cache to train() entry - speeds up processing by ~4x if you have a large number of similar values (e.g. a boolean field with Y/N)
• BUG: NULL_TEXT_AS_NULL not being honored in trainBulk()
• BUG: Fix NPE in isValid() for some addresses
• INT: Bump org.springframework.boot to 3.4.1, com.github.ben-manes.versions to 0.52.0, gradle to 8.12.1, google phonenumber to 8.13.55, testng to 7.11.0
• ENH: Enhance documentation as it relates to sampleCount and merging

15.12.0

• ENH: Add new Semantic Types - SPATIAL.WKT, SPATIAL.GEOJSON (#121)
• INT: Bump logback-classic to 1.5.16, google phonenumber to 8.13.53
• INT: Some minor cleaning
• BUG: If FORMAT_DETECTION is enabled - error messages were being printed to stdout