Scoring Engine

This section covers the scoring engine installation, configuration and start-up.

Installation

The scoring engine repositories are automatically installed as part of the ATK repositories.

Scoring Models Implementation

The scoring engine is independent of the streaming scoring model implementation. To obtain information concerning the model implementation, the scoring engine expects three files in a tar file:

1. The model implementation jar file.
2. A file modelname.txt that contains the name of the class that implements the scoring in the jar file.
3. The file that has the model bytes used by the scoring.

Note

If Trusted Analytics Platform is used to build models, the publish method on the model will create the tar file needed by the scoring engine.

Configuration of the Engine

The scoring engine provides a configuration template file application.conf.tpl which is used to create a working configuration file application.conf. Copy the configuration template file to the working file name in the same folder:

$ cd /etc/trustedanalytics/scoring
$ sudo cp application.conf.tpl application.conf

Open the file with a text editor:

$ sudo vim application.conf

Modify the scoring engine section to indicate where the scoring tar file is located:

trustedanalytics.scoring-engine {
  archive-tar = "hdfs://scoring-server.company.com:8020/user/atkuser/kmeans.tar"
}

Starting the Scoring Engine Service

Once the application.conf file has been modified, the scoring engine can be started:

$ sudo service scoring-engine start

Launch the rest server for the engine:

GET /v1/models/[name]?data=[urlencoded record 1]

See the REST API for more information.

Scoring Client

An example python script to connect to the scoring engine:

>>> import requests
>>> import json
>>> headers = {'Content-type': 'application/json',
...            'Accept': 'application/json,text/plain'}
>>> r = requests.post('http://localhost:9100/v1/models/testjson?data=2', headers=headers)