rivethead/ds-workbench

Workbench

Getting started

• Run just create-virtual-env to create the Python virtual environment
• Run just install to set up the virtual environment and install requirements
• Run just delete-virtual-env to remove the Python virtual environment

Infrastructure

The workbench infrastructure is inspired and informed by https://emilygorcenski.com/post/configuring-a-data-science-workbench/

• Run just env-up to:

  • Start MLflow server - http://localhost:5555
  • Start PostgreSQL database
  • Start MinIO - http://localhost:9000
  • Start JupyterLab - http://localhost:8888/?token=neely

• Run just env-down to:

  • Stop MLflow server
  • Stop PostgreSQL database
  • Stop MinIO

• Passwords in default.env file

Bench

Wine example

• Navigate to the bench/wine directory and run just
• This will execute the train.py script and push the results to MLflow
• The .env controls the environment variables indicating where MLflow is running