Kamanja v1.6 – released Sep 30, 2016
There is a new release of Kamanja – Version 1.6 – released Sep 30, 2016!
- Source code can be found at https://github.com/LigaData/Kamanja/releases.
- Download and install using RPM. Find the links to the RPM 1.6.0:
- https://s3.amazonaws.com/kamanja/release/2.11/Kamanja-1.6.0_2.11.rpm (choose if Scala 2.11 is installed).
- https://s3.amazonaws.com/kamanja/release/2.10/Kamanja-1.6.0_2.10.rpm (choose if Scala 2.10 is installed).
- Kamanja binaries are available at:
- https://s3.amazonaws.com/kamanja/release/2.11/Kamanja-1.6.0_2.11.tar.gz (choose if Scala 2.11 is installed).
- https://s3.amazonaws.com/kamanja/release/2.10/Kamanja-1.6.0_2.10.tar.gz (choose if Scala 2.10 is installed).
This major release introduces the following new features:
- Model Support
- Python Support Added - provides built-in support for Python 2.7 within the engine. This allows developers and data scientists to leverage their existing Python models and preprocessing logic without having to rewrite any code. See here for details.
- File Adapter - improves file adapter by providing input directory scaling to be able to have different sizes of input files. Enables selective processing of files that allows the client to specify what type of file format to read. See here for details.
- Ease of Use
- New Sample Models/Tutorials - validates models to enable data scientists to compare Kamanja performance with other commercial products. See here for details.
- RPM Replaces Homebrew - provides ease of use, upgradability of packages, single command installation, and package verification. See here for details.
- 1.6 Migration - enhances cluster upgrade to allow users to migrate to Kamanja 1.6.0. The cluster upgrade instructions are the same as in previous versions except that the version number is now 1.6.0. See here for details.
- Bug Fixes
- GitHub Issue 998, 999, 1006, 1036 - improves code consistency so that all metadata operations(CRUD) return API Results.
- GitHub Issue 1159 - enhances Kamanja engine so that it takes in custom engine properties as arguments. Before the user had to tediously modify the default properties file but now he/she can simply provide a custom properties file.
- GitHub Issue 805 - upgrades code so that the MetadataAPI and KamanjaManager components accept JVM options as part of the configuration. Now the user can provide JVM options such as memory capacity through either the properties files or input arguments.