Python Images

Includes Docker images for ML and batch operations using Python

Running Jobs

Package	Definition
py-dynamic:dynamic	Base container for Python based jobs, installing a given Python library and running a specific module on start-up
py-dynamic:dynamic-media	Dynamic container with additional dependencies (e.g. ffmpeg) installed for py-media operations
py-dynamic:dynamic-java	Dynamic container with JDK installed for Java dependent operations (e.g. Spark)
py-dynamic:dynamic-java-hadoop	Dynamic container with Hadoop installed for file system dependent operations

Additional versions of images are also applicable, combining media, java and hadoop (e.g. py-dynamic:dynamic-media-java-hadoop).

These images are accessible from "rierino-open" Docker repository and have the following environment variables used in their default command:

Variable	Description	Example
GITHUB_USER	GitHub user to use for accessing Rierino Python packages	-
GITHUB_AUTH_TOKEN	GitHub token to use for accessing Rierino Python packages	-
PY_REPO	Repository for loading main Rierino Python module	rierino-open/py-runner
PY_RIERINO_VERSION	Version of main Rierino Python module to load	0.1.1
PY_SCRIPT	Additional script to execute (e.g. for loading extra libraries)	/app/config/script.sh
PY_MAIN_MODULE	Main module & arguments to pas on to main Rierino Python module for execution	rierino_util.Runner --base64 XXX

These environment variables are configured through Helm Chart parameters when the image is deployed using Batch Charts.

Loader Jobs

Package	Definition
py-loader:loader	Base container for Python based deployment assets loader, used during initial installations and version upgrades
py-loader:loader-core-train	Loader container with core and training deployment assets pre-downloaded

These images are accessible from "rierino-open" Docker repository and are typically used with Helm Charts during installation and version upgrades.