Skip to content

Latest commit

 

History

History
58 lines (39 loc) · 2.88 KB

README.md

File metadata and controls

58 lines (39 loc) · 2.88 KB

Travis

PredictionIO chef cookbook

PredictionIO is an open source Machine Learning Server built on top of state-of-the-art open source stack for developers and data scientists create predictive engines for any machine learning task. PredictionIO cookbook aims to provide two variants of installation standalone and all-in-one (AIO).

The standalone variant installs and configures a system to run PIO EventServer/PredictionServer or a PIO driver machines. This installs the PIO stack sources (Hadoop/HBase/Spark), builds PIO and writes its configuration files.

The second variant is AIO installs and sets the system up with all the required services on a single machine. Apache dependencies such as Hadoop/HBase/Spark are configured in Pseudo Distributed mode. It's strongly suggested to the end-user to use this variant to evaluate PredictionIO.

Current software versions

  • PredictionIO - 0.12.1
  • Universal Recommender - 0.7.3
  • OpenJDK8
  • Hadoop - 2.8.4
  • HBase - 1.4.4
  • Spark - 2.1.2
  • ElasticSearch - 5.X

We support:

  • Debian >= 8
  • Ubuntu >= 14.04
  • CentOS >= 7

Actually any systemd/upstart based system should work smooth, though there are cookbook metadata restrictions for the given platforms. Feel free to change this, please open a PR and contribute.

Cookbook attributes

This cookbooks contains quite a long list of attributes, we won't try to cover all of them here rather than just try to introduce the most important ones to you. For the complete list see attributes/.

Attribute Description Default value
pio.user User used for operating PredictionIO. 'aml'
pio.home PIO user's home directory. Specify to change the default (/opt/data/aml-home which is symlinked to /home/aml). ''
pio.localdir Directory containing all the installed applications and services. '/usr/local'
pio.datadir Directory used as the top data store for all the installed services. '/opt/data'
pio.pio.install_method Specifies PredictionIO installation method (either binary source) 'source'
pio.ur.git* Attributes setting the Universal Recommender Git source location and version.
pio.conf.event_server Specifies the PredictionIO Event Server port. 7070
pio.conf.prediction_server Specifies the PredictionIO Prediction Server port. 8000
pio.hdfs_structure Specifies initial HDFS layout to bootstrap. see attributes

Usage

pio::default

For the standalone installation, include pio recipe in your node's run_list.

pio::aio

For the AIO installation, include pio::aio recipe in your node's run_list.

License and Authors