rsolano/python-notebook repository overview

⁠Python Jupyter Notebook

A Debian⁠-based personal Jupyter Python workstation. Provides Python + basic engineering modules, iPython and Jupyter Notebook.

Ramon Solano <[email protected]⁠>

Last update: May/22/2019
Debian version: 9.9

⁠Main packages

• Python[2,3]
  • Modules: Numpy, Pandas, SciPy
  • Plotting: Matplotlib, Seaborn, Plotly
• IPython
• Jupyter Notebook

⁠Users

User/pwd:

• root/debian
• debian/debian (sudoer)

⁠To build from Dockerfile

$ docker build -t rsolano/python-notebook .

⁠To run container

$ docker run [-it] [--rm] [--detach] [-h HOSTNAME] [-p LNOTEBOOKPORT:8888] [-v LDIR:DIR] rsolano/python-notebook [cmd]

where:

• LNOTEBOOKPORT: Local HTTP Jupyter Notebook port to connecto to (e.g. 8888). Requires IP=0.0.0.0 when running Jupyter in your container for connecting from your locahost, otherwise IP=127.0.0.1 for internal access only.

• LDIR:DIR: Local directory to mount on container. LDIR is the local directory to export; DIR is the target dir on the container. Both sholud be specified as absolute paths. For example: -v $HOME/worskpace:/home/debian/workspace.

• cmd: User command to run inside the container

⁠Examples

• Run image, keep terminal open (interactive terminal session); remove container from memory once exited from container; map Jupyter Notebooks to 8888; mount local $HOME/workspace on container's /home/debian/workspace:

  $ docker run -it --rm -p 8888:8888 -v $HOME/workspace:/home/debian/workspace rsolano/python-notebook
  

  Copy

• As above, but running immediately Jupyter:

  $ docker run -it --rm -p 8888:8888 -v $HOME/workspace:/home/debian/workspace rsolano/python-notebook jupyter-notebook --ip 0.0.0.0 --no-browser
  

  Copy

• Run image, detach to background and keep running in memory (control returns to user immediately, Notebook keep running in background); map Jupyter Notebook to 8888; mount local $HOME/workspace on container's /home/debian/workspace:

  $ docker run --detach -p 8888:8888 -v $HOME/workspace:/home/debian/workspace rsolano/python-notebook jupyter-notebook --ip 0.0.0.0 --no-browser
  

  Copy

NOTES

• There is no graphical desktop; just the Jupyter Notebook engine to serve Notebooks

• Mapping Jupyter container port 8888 to our host's port 8888, allows us to connect to our computer (localhost:8888), which is actually forwarding to the container.

• To run Jupyter in a headless container (no graphical display), you shoukd specify:

  • Option --no-browser : Do not open a local web browser
  • Option --ip 0.0.0.0 : Allow connecting to Jupyter server from external computers

  This means, run jupyter as:

  $ jupyter-notebook --ip 0.0.0.0 --no-browser
  

  Copy

⁠To stop container

• If running an Jupyter in an interactive session:

  • Just press CTRL-C in the interactive terminal to quit Jupyter.
  • Just press CTRL-D or key in exit end finish the session.

• If running detached (background) session:

  1. Look for the container Id with docker ps:

     $ docker ps
     CONTAINER ID        IMAGE                     COMMAND                  CREATED             STATUS              PORTS                                          NAMES
     ac46f0cf41d1        rsolano/python-notebook   "/usr/bin/supervisor…"   58 seconds ago      Up 57 seconds       0.0.0.0:5900->5900/tcp, 0.0.0.0:2222->22/tcp   wizardly_bohr
     

     Copy

  2. Stop the desired container Id (ac46f0cf41d1 in this case):

     $ docker stop ac46f0cf41d1
     

     Copy