pynstal

A CLI utility to manage Python virtual environments with auto-detection of system Python installations.

Features

• Auto-detect Python installations across Windows, macOS, and Linux (pyenv, conda, system Python)
• Manage interpreters - add/list/remove and set defaults
• Project-aware defaults - per-project pynstal.json plus global config under ~/.pynstal
• Create virtual environments using any configured Python interpreter
• Install templates - predefined and custom package templates (including pip args like CUDA wheels)
• Persistent configuration - interpreter paths and templates stored in global config, with optional project-local defaults

Installation

From PyPI

pip install pynstal

After installation, use the pynstal command:

pynstal --help

From source

git clone https://github.com/yourusername/PythonProjectManager.git
cd PythonProjectManager
pip install -e .

Quick Start

1. Detect and add Python installations

# List found Python installations
pynstal interpreter detect

# Interactively select which to add
pynstal interpreter detect --add

# Add all detected interpreters
pynstal interpreter detect --add-all

2. Set a default interpreter (optional)

pynstal set-default-interpreter   # prompts to choose from the configured list

3. Create a virtual environment

# Uses project-local default (pynstal.json) or global default if not provided
pynstal create-venv my_venv

# Or specify an interpreter explicitly
pynstal create-venv my_venv --interpreter "/usr/bin/python3.12"

4. Install packages using templates

# List available templates
pynstal template list

# Create venv with datascience template (numpy, scipy, pandas, etc.)
pynstal create-from-template datascience my_ds_venv

# Create venv with PyTorch (CUDA 12.4)
pynstal create-from-template pytorch-cu124 my_pytorch_venv

5. Create custom templates

# Create a template (interactive)
pynstal template add mytemplate
# then add packages
pynstal template add-pkg mytemplate numpy scipy matplotlib
# or add packages with custom pip args
pynstal template add-pkg-complex mytemplate torch torchvision --args-str "--index-url https://download.pytorch.org/whl/cu124"

# Create venv from the custom template
pynstal create-from-template mytemplate my_venv

Commands

• Interpreter management

  • pynstal interpreter list - List configured interpreters
  • pynstal interpreter detect [--add | --add-all] - Auto-detect system Python installations
  • pynstal interpreter add - Interactive prompt to add interpreter paths
  • pynstal interpreter remove - Interactive prompt to remove interpreter paths
  • pynstal add-interpreter <path> - Manually add a Python interpreter path (positional)
  • pynstal set-default-interpreter - Interactive prompt to choose the global default interpreter

• pynstal create-venv <venv_dir> [--interpreter <path>] [--dry-run] - Create a virtual environment

• pynstal create-from-template <template> <venv_dir> [--interpreter <path>] [--dry-run] - Create venv and install template packages

• pynstal remove-venv <venv_dir> - Delete a virtual environment and clean up references

• pynstal install <template> [--interpreter <path>] [--dry-run] - Install packages from a template into an interpreter without creating a venv

• Template management

  • pynstal template list - List all templates
  • pynstal template show <name> - Show template details
  • pynstal template create <name> - Create a new template interactively
  • pynstal template add-pkg <name> <packages...> - Add packages to an existing template
  • pynstal template add-pkg-complex <name> <packages...> --args-str "<args>" - Add packages with pip args to a template
  • pynstal template remove-pkg <name> - Interactively remove packages from a template
  • pynstal template remove <name> - Remove a template

Data Storage

• Global config: ~/.pynstal (or $XDG_CONFIG_HOME/pynstal)
  • interpreters.json – configured interpreter paths and default_interpreter
  • templates.json – user-defined package templates
• Project-local config: ./pynstal.json
  • default_interpreter – default interpreter for the current project (set automatically when creating a venv)

Platform Support

• Windows: Detects pyenv-win, AppData Python installations, system Python
• macOS/Linux: Detects /usr/bin/python*, pyenv, conda, system Python

License

MIT

Contributing

Contributions welcome! Please feel free to submit issues and pull requests.