DSCO

A utility to create a data science development container.

Documentation

github pages

Features

DSCO creates a development container that:

• runs a jupyter notebook server.
• generates static html representations of your notebooks for easy sharing with anyone by either the included server or github pages.
• manages python dependencies with poetry.
• integrates with vscode for an IDE with a local-like feel.

If that's not enough, you also get:

• Flask and plotly | Dash for creating dashboards and REST APIs.
• Sphinx for automatic code documentation.
• Useful Jupyter extensions installed and configured.
• Your local project directory mounted and synced.
• Ansible playbooks that build everything from scratch.

Install

pip install git+ssh://git@github.com/Teradata/dsco.git

Requirements:

• pip > version 10
• docker: tested with version 2.2.0.4
• python >= 3.7

Ultra Quickstart

See also: Quickstart

Create project foo in the current directory:

>>> dsco init
Using template: ...
Creating project in ...
project_name [MyProject]: foo
author [Discovery]:
version [0.1]:
year [2020]:
project_port [8001]:
select python_version:
1 - 3.7
2 - 3.6
Choose from 1, 2 [1]:

Build the development image and create a development container:

>>> cd foo
>>> dsco go
docker-compose up -d dev
Building dev
...
<lots of ansible output>
...
Creating foo_dev_1 ... done
========================================================================================
execution time: 490.9753198623657 seconds
========================================================================================