/deep-start

Startup script for DEEP containers

Primary LanguageShellMIT LicenseMIT

Start script(s) for DEEP/AI4OS containers

deep-start

The main start script for DEEP/AI4OS containers. Automatically detects if NVIDIA-GPU is present in the system. Default start is equivalent to deep-start -d.

Usage: deep-start

Options:
-h|--help         the help message
-c|--cpu          force CPU-only execuition (otherwise detected automatically)
-g|--gpu          force GPU execution mode (otherwise detected automatically)
-d|--deepaas      start deepaas-run
-i|--install      enforce that the latest git repo of the deep-start script is installed
-j|--jupyter      start JupyterLab; if not installed, will be automatically installed
-o|--onedata      mount remote storage using oneclient
-s|--vscode       start VSCode (code-server); if not installed, will be automatically installed
-v|--version      print script version and exit

NOTE: if you try to start deepaas-run AND jupyterlab (or vscode), only deepaas-run will start!

jupyter_notebook_config.py

(and symlink jupyter_server_config.py)

Module to set Jupyter access password from the jupyterPASSWORD environment, if available. Tested for jupyterlab v2,3,4. In addition it provides default settings for JupyterLab, e.g.:

  • --ip
  • --port
  • --no-browser
  • --allow-root

(directory) lab

contains very basic configuration for Jupyter Lab

(directory) vscode

contains very basic configuration for VSCode

  • vscode/code-server/vscode-extensions.txt : list of default extensions installed for VSCode

During execution of deep-start a remotly located at github vscode-extensions.txt is probed to load.

!! Depricated

run_jupyter.sh : script to start jupyterlab =>> now deep-start starts jupyterlab directly