Skip to main content

Notebook

Recipe Features Explained​

Let's breakdown the recipe to deploy a new Jupyter Notebook site.

See Full Deployment Script 
source: clear
features:
  - python stable
nginx:
  root: public_html/public
  passenger:
    enabled: "on"
    app_start_command: jupyter notebook --port $PORT
commands:
  - pip install jupyter
  - mkdir -p ~/.jupyter ; cd $_ ; rm -f jupyter_notebook_config.py ; jupyter notebook --generate-config
  - hash=`python -c "from jupyter_server.auth import passwd;print(passwd('${PASSWORD}'),end='')"`
  - sed -i "s@# c.ServerApp.password = ''@c.ServerApp.password = u'${hash}'@g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.password_required = False/c.ServerApp.password_required = True/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.quit_button = True/c.ServerApp.quit_button = False/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.open_browser = True/c.ServerApp.open_browser = False/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.allow_origin = ''/c.ServerApp.allow_origin = '*'/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.allow_remote_access = False/c.ServerApp.allow_remote_access = True/g" jupyter_notebook_config.py
  - pip install pandas numpy scipy matplotlib scikit-learn seaborn
source: clear
features:
  - python stable

This script clears ~/public_html and installs a new Python via Pyenv.

nginx:
  root: public_html/public
  passenger:
    enabled: "on"
    app_start_command: jupyter notebook --port $PORT

This script declare NGINX Configuration. It means two thing:

  1. root: public_html/public means all static files are put to public folder. This means nothing but important so your python files don't accidentally exposed.

  2. app_start_command: jupyter notebook --port $PORT means when you visit the site, it run jupyter notebook --port $PORT. The $PORT assignment is important to know which port that NGINX forward requests to.

commands:
  - pip install jupyter

This script installs jupyter notebook via pip.

commands:
  - mkdir -p ~/.jupyter ; cd $_ ; rm -f jupyter_notebook_config.py ; jupyter notebook --generate-config
  - hash=`python -c "from jupyter_server.auth import passwd;print(passwd('${PASSWORD}'),end='')"`
  - sed -i "s@# c.ServerApp.password = ''@c.ServerApp.password = u'${hash}'@g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.password_required = False/c.ServerApp.password_required = True/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.quit_button = True/c.ServerApp.quit_button = False/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.open_browser = True/c.ServerApp.open_browser = False/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.allow_origin = ''/c.ServerApp.allow_origin = '*'/g" jupyter_notebook_config.py
  - sed -i "s/# c.ServerApp.allow_remote_access = False/c.ServerApp.allow_remote_access = True/g" jupyter_notebook_config.py

This script:

  1. run jupyter notebook --generate-config to generate config to ~/.jupyter
  2. Installs jupyter password as $PASSWORD. This variable is a server password found from Manage tab.
  3. Set jupyter config:
  • c.ServerApp.password_required = True
  • c.ServerApp.quit_button = False
  • c.ServerApp.open_browser = False
  • c.ServerApp.allow_origin = '*'
  • c.ServerApp.allow_remote_access = True
commands:
  - pip install pandas numpy scipy matplotlib scikit-learn seaborn

Installs popular depedencies to run jupyter notebook:

  • pandas
  • numpy
  • scipy
  • matplotlib
  • scikit-learn
  • sciborn

TO DO​

TODO