ColCarroll / callisto

Callisto

The fourth Galilean moon of Jupyter.

A command line utility to create kernels in Jupyter from virtual environments.

Installation

Callisto may be installed from pypi:

pip install callisto

Tested against python 2.7, 3.4, 3.5, 3.6.

Basic Usage.

Typical use is to just activate it inside a virtual environment:

$  virtualenv venv
...
$  source venv/bin/activate
(venv) $  callisto
Successfully installed a new jupyter kernel "venv":
{
  "env": {},
  "language": "python",
  "display_name": "venv",
  "argv": [
    "/Users/colin/venv/bin/python",
    "-m",
    "ipykernel",
    "-f",
    "{connection_file}"
  ]
}
See /Users/colin/Library/Jupyter/kernels/venv/kernel.json to edit.

Jupyter servers will now have an option for a kernel called venv.

Naming the kernel.

You may also give kernels a more descriptive name:

(venv) $  callisto -n pete
Successfully installed a new jupyter kernel "pete":
{
  "env": {},
  "display_name": "pete",
  "argv": [
    "/Users/colin/venv/bin/python",
    "-m",
    "ipykernel",
    "-f",
    "{connection_file}"
  ],
  "language": "python"
}
See /Users/colin/Library/Jupyter/kernels/pete/kernel.json to edit.

Jupyter servers will now have an option for a kernel called venv, and pete.

Deleting kernels.

Sometimes you may want to tidy kernels up a bit.

(venv) $  callisto -d
Deleted jupyter kernel "venv" from /Users/colin/Library/Jupyter/kernels/venv/kernel.json:
{
  "argv": [
    "/Users/colin/venv/bin/python",
    "-m",
    "ipykernel",
    "-f",
    "{connection_file}"
  ],
  "env": {},
  "language": "python",
  "display_name": "venv"
}

Jupyter servers will no longer have a kernel named venv.

Lacking courage.

Callisto doesn't try to be too clever.

(venv) $  deactivate

$  callisto
Usage: callisto [OPTIONS]

Error: The environment variable VIRTUAL_ENV is not set (usually this is set
automatically activating a virtualenv).  Please make sure you are in a
virtual environment!

Viewing existing kernels.

If you forgot the informative message about the kernel information, you can see it later.

$  source venv/bin/activate

(venv) $  callisto --list
No kernel found at /Users/colin/Library/Jupyter/kernels/venv/kernel.json

(venv) $  callisto -l --name pete
Found kernel "pete" at /Users/colin/Library/Jupyter/kernels/pete/kernel.json:
{
  "display_name": "pete",
  "language": "python",
  "argv": [
    "/Users/colin/venv/bin/python",
    "-m",
    "ipykernel",
    "-f",
    "{connection_file}"
  ],
  "env": {}
}

Adjusting the PYTHONPATH.

With isolated kernels, you may wish to run all your notebooks from a single directory, but using code from the project directories.

(venv) $  callisto -n pete --path=$(pwd)
Successfully installed a new jupyter kernel "pete":

{
  "argv": [
    "/Users/colin/venv/bin/python",
    "-m",
    "ipykernel",
    "-f",
    "{connection_file}"
  ],
  "language": "python",
  "env": {
    "PYTHONPATH": "/Users/colin/projects/pete:PYTHONPATH"
  },
  "display_name": "pete"
}
See /Users/colin/Library/Jupyter/kernels/pete/kernel.json to edit.

Now the pete kernel will be able to import from the folder /Users/colin/projects/pete.