Adding Python 3 libraries

Version 5.0.0 and higher of the Machine Learning Toolkit (MLTK) requires version 2.0 of the Python for Scientific Computing add-on, version 8.0 of Splunk Enterprise, and Python 3.

Users on this version or above of the Machine Learning Toolkit have the option to add Python 3 libraries as a means to enhance their machine learning efforts.

Support is not offered on the use of or upgrade of any Python 3 libraries added to your Splunk platform instance. Any upgrade to MLTK or the PSC add-on will overwrite any Python library changes.

Follow these steps to add a Python 3 library to your instance of the MLTK:

1. Clone the GitHub repo for the Python for Scientific Computing add-on: https://github.com/splunk/Splunk-python-for-scientific-computing.git
2. Navigate to https://repo.anaconda.com/pkgs/ to check the list of packages supported through Anaconda. You can only add packages listed on this site.
3. In GitHub, choose the package you need and add it in package.txt.
4. Specify the version of the package in package.txt. The latest version is selected by default.
5. Run bash repack.sh to create the environment and install the package within the environment.
6. When the repacking in complete, run the bash build.sh script which creates a .tgz file for the PSC add-on. On Windows, run a build.psl script.
7. In your Splunk platform instance (not in the Splunk CLI or web installer) extract the .tgz file.

The final .tgz app stores in the build directory.