Release 2.1.0

What's Changed

• Add sample weight to LightGBM by @jpleitao in #147
• Bump maven-javadoc-plugin version by @artpdr in #148
• Use Portal OSSRH Staging API by @artpdr in #149
• fix(lightgbm): ensuring c++ resources created in swig are released after training by @jpleitao in #150
• Give id token permissions to fix codecov step in master by @artpdr in #151

New Contributors

• @jpleitao made their first contribution in #147

Full Changelog: 2.0.1...2.1.0

Release 2.0.1

What's Changed

• Add support for Alpine on LGBM by @Renato2000 in #145

New Contributors

• @Renato2000 made their first contribution in #145

Full Changelog: 2.0.0...2.0.1

Release 2.0.0

What's Changed

• Remove distribution of fat-jars with all dependencies by @TiagoMSSantos in #143

Note

This new major release is potentially a breaking change since the projects using feedzai-openml-java might need to add some dependencies in the project that feedzai-openml-java depends on, or may need to exclude some transitive dependencies from feedzai-openml-java if it is already using a different version.

Full Changelog: 1.6.0...2.0.0

Release 1.6.0

What's Changed

• Fixed pinned FairGBM tag to v0.1.0 by @AndreFCruz in #137
• [fix] make-lightgbm docker build by @AlbertoEAF in #140
• Add support for ARM64 on LGBM by @artpdr in #142

New Contributors

• @artpdr made their first contribution in #142

Full Changelog: 1.5.0...1.6.0

1.5.0

What's Changed

• Reverted change to learning_rate that made it mandatory by @antssilva96 in #135
• Document experimental releases in README by @antssilva96 in #136

Note

This release should be used instead of 1.4.X.

When doing an experimental release for 1.4.1 a problem occurred and an actual release was done with version 1.4.9999.

1.4.0

New features

• FairGBM openml API implementation by @AndreFCruz in #116
• updated since tags to version 1.4.0 by @AndreFCruz in #133
• Pin FairGBM artifact version to fairgbm-v1.0.0 by @AndreFCruz in #134

Maintenance

• Bump jetty and log4j in H2O provider to avoid CVEs by @paulojrp in #102
• Fix validation when training H2o models by @paulojrp in #103
• Fix unrecognized format specifier by @paulojrp in #110
• Fix unrecognized format specifier when training xgboost models by @paulojrp in #111
• Bump gson from 2.6.2 to 2.8.9 by @dependabot in #115
• Stop ignoring LightGBMBinaryClassificationModelTrainerTest by @mlobofeedzai in #119
• ref PULSEDEV-36792 openml-lightgbm: Fix segmentation fault by @mlobofeedzai in #124

Release 1.2.1

Maintenance

• Update make-lightgbm submodule, which solves deprecated ca-certificates issue on docker build.
• Bump commons-io from 2.4 to 2.7
• Bump httpclient from 4.5.5 to 4.5.13, which avoid duplicate class issue.
• Bump from log4j to log4j-core

Release 1.1.4

Maintenance

• Update make-lightgbm submodule, which solves deprecated ca-certificates issue on docker build.
• Bump commons-io from 2.4 to 2.7
• Bump httpclient from 4.5.5 to 4.5.13, which avoid duplicate class issue.

Release 1.0.20

Maintenance

• Update make-lightgbm submodule, which solves deprecated ca-certificates issue on docker build.
• Bump commons-io from 2.4 to 2.7
• Bump httpclient from 4.5.5 to 4.5.13, which avoid duplicate class issue.

Release 1.2.0

New Features

• openml-lightgbm: Add LightGBM streaming data support (#55)
  Reduces LGBM model train memory consumption by 50-70%.

Maintenance

• pom: Bump log4j to version 1.2.17-cloudera1 (#59)
• Allow snapshot publish