mogger: Logger for Predictive Models

The application that allows storing the metadata of models and their audits in the database. It also provides a web interface with visualizations of the database contents. Storing and acquiring data is made possible through the REST API.

Components

• mogger - JAVA REST application (server with database)
• docker - docker file that allows to run this service on any machine
• docs - documentation with all the information
• python_package - Python client package to work with the service
• r_package - R client package to work with the service
• scripts - other useful files e.g. script that creates the database
• examples - Python/R examples

Run

Build a docker by typing in a mogger/docker directory:

sudo docker build -t mogger .

Run the container with:

sudo docker run mogger

Then by:

sudo docker inspect $id

where $id is your docker's id, you can check the app ip address.

Default $id is 172.17.0.2. Docker $id is a hash from:

sudo docker ps

Example