Docker Images for Common Dev/Production Environment

There are 3 docker images available: python2, python2-teradata, python3, python3-teradata, jupyter-python-ultra-pack.

python2

docker pull raychee/python2

This image is based on CentOS 7, bundled with miniconda (the minimized version of Anaconda). That means python (version 2.7) and pip are already available. conda is also available of course, and it is recommended for package management over pip. To learn more about conda, please check out its homepage.

• docker run -it raychee/python2
  enters the bash shell in CentOS 7 and you are good to go.

• docker run --rm -v `pwd`:/src raychee/python2 python /src/<your_script_in_local_dir>.py
  can run a script in your current local directory, with python in the docker environment.

python2-teradata

docker pull raychee/python2-teradata

This image is based on python2, with additional Teradata ODBC driver installed. Specifically, the driver includes Teradata Linux Driver, and teradata the python package.

Here is an example code snippet for connecting to Teradata with python (run in docker):

import teradata
uda = teradata.UdaExec(appName='whatever', version='1.0')
session = uda.connect(method='odbc', system='<www.host_server.com>', username='<your_name>', password='<your_password>')
query = session.execute('SELECT * FROM table')
result = query.fetchall()

(There is an even simpler solution for connecting to Teradata, but for python 3 only: here)

python2-teradata-ldap

docker pull rexren/python2-teradata-ldap

This image is based on python2-teradata, with additional ldap related packages installed. You can type import ldap to use python-ldap.

python3

docker pull raychee/python3

This image is identical to python2 except that the python version is 3.6.

python3-teradata

docker pull raychee/python3-teradata

This image is identical to python2-teradata except that the python version is 3.6.

By the way, there is an even simpler and more friendly solution to connect to teradata (for python 3 only): firstly pip install qutils in bash, and then in python

from qutils.iotools import Teradata
tera = Teradata('<www.host_server.com>', '<your_user_name>', '<your_password>')
result = tera.query('SELECT * FROM table')

This result will be a Pandas DataFrame.

jupyter-python-ultra-pack

docker pull raychee/jupyter-python-ultra-pack

This is an ALL-in-ONE image based on python3-teradata. It includes python 2.7, python 3.6, Anaconda, Teradata Linux Driver, xgboost (0.6), and a pre-configured jupyter notebook.

• docker run -d -p 8888:8888 -v `pwd`:/jupyter raychee/jupyter-python-ultra-pack
  starts a jupyter server in local host which will be listening on all open IP login requests.
  For example, you can deploy it on C3 and open http://<your_c3_host>:8888 with your local computer to run any python code on C3.
  The python 2 and 3 kernels are already available; you can feel free to switch it.

Feedback

Please feel free to try it and let me know any issues when you are using the images.