.local-bin/docs/add_ipykernel/Usage.md

add_ipykernel

Adds Python virtual environments as Jupyter ipykernel options.

Description

This script automates the creation and registration of Python environments as Jupyter kernels. It supports both conda and pip virtual environments.

Usage

./add_ipykernel.sh ENV_NAME [--create] [--verbose] [pkg1 pkg2 ...]

Arguments

• ENV_NAME: Name of the environment/kernel to create or use
• --create: Create a new environment if it doesn't exist
• --verbose: Enable detailed output
• pkg1 pkg2 ...: Additional packages to install

Examples

# Create new conda environment and register as kernel
./add_ipykernel.sh ds_env --create python=3.11 pandas

# Use existing environment
./add_ipykernel.sh ds_env --verbose

# Create pip venv with specific packages
./add_ipykernel.sh myproject --create numpy scipy matplotlib

Package Manager Detection

The script automatically detects the appropriate package manager:

• conda: Used if environment exists in conda or if environment.yml/conda.yaml is present
• pip: Used if requirements.txt or pyproject.toml exists, or if already in a virtual environment

What It Does

1. Detects package manager (conda or pip)
2. Creates environment if --create flag is used
3. Activates the environment
4. Installs ipykernel if not present
5. Registers the environment as a Jupyter kernel with display name Python (ENV_NAME)

Requirements

• conda (for conda environments) OR python3 with venv (for pip environments)
• Jupyter installed (for kernel registration)

Notes

• Kernels are registered in user space (--user flag)
• After registration, the kernel appears in Jupyter as "Python (ENV_NAME)"
• If environment already exists, just registers it without modifying packages