This Dockerfile installs various slow-to-install dependencies, like NumPy and SciPy.
It's used as a base-image for other Python images. By separating this slow, infrequently-changing part of the build, automated Docker builds are considerably faster.
GUnicorn, Flask, and PsycoPG are included because the final images present the analytical code as a REST service.
- Python 3
- NumPy 1.8
- SciPy 0.13
- Flask 0.10
- GUnicorn 19
- python-dateutil 2.2
- pgpasslib
- psycopg2