Getting Started with OmegaML

OmegaML is an API service that consists of a worker connected to a broker (rabbitmq) and a mongo database. It enables an end user do offload all the heavylifting involved with running analytics.

Installation

OmegaML can be easily installed with pip using

sudo pip install omegaml

Configuration

OmegaML takes much of its configuration parameters from omegaml.defaults. For any customization all you need to do is export an environment variable with the same name. For e.g, say if you want to change the mongo url to some destination other than the default on your shell all you need to do is

# export OMEGA_MONGO_URL='user@host:port/db

and omegaml reconfigure itself. See omegaml.defaults for more.

Using OmegaML on private Spark Cluters

OmegaML can be installed on any machine of your preference following below steps:

  • install omegaml using pip

  • set env variables for mongodb and rabbitmq

  • start celery worker

  • import omegaml on start of pyspark