Python Environment: Miniforge

Miniforge is a powerful package management system that provides fast and reliable installation of scientific packages and their dependencies.

What is Miniforge?

Miniforge is a distribution of the conda package manager that uses the fast mamba solver instead of the traditional conda solver. It includes:

  • Mamba: A fast, cross-platform package manager

  • Conda: The classic conda package manager

  • Python: A Python interpreter

  • Essential packages: numpy, pandas, scipy, etc.

Why Choose Miniforge?

  • Speed: Mamba is significantly faster than conda (often 10-100x faster)

  • Reliability: Better dependency resolution

  • Cross-platform: Works on Windows, macOS, and Linux

  • Large ecosystem: Access to millions of packages via conda-forge

  • Environment management: Create isolated environments for different projects

Installation

Download Miniforge

  1. Visit the official download page: https://github.com/conda-forge/miniforge

  2. Choose the installer for your operating system:

    • Windows: Download Miniforge-Windows-x86_64.exe

    • macOS: Download Miniforge-MacOSX-x86_64.sh (Intel) or Miniforge-MacOSX-arm64.sh (Apple Silicon)

    • Linux (WSL): Download Miniforge-Linux-x86_64.sh

Install on Windows

  1. Run the downloaded .exe file

  2. Follow the installation wizard

  3. Choose installation location (recommended: C:\Users\<username>\miniforge3)

  4. Check “Add Miniforge to PATH” (optional, but recommended for beginners)

  5. Complete installation

Note

I recommend using Windows Subsystem for Linux (WSL) for a more Unix-like experience when working with Miniforge on Windows. Please see the WSL setup guide for more details.

Install on macOS/Linux (including WSL)

  1. Open Terminal

  2. Navigate to download directory:

    cd path/to/downloads

  3. Run the installer script:

    sh Miniforge-*.sh

  4. Follow the prompts:

    • Press Enter to read the license

    • Type yes to accept

    • Choose installation location (default is recommended)

    • Type yes to initialize Miniforge

  5. Restart your terminal or run:

    exec $SHELL

Verify Installation

After installation, verify it works:

mamba --version
conda --version

You should see version numbers for both mamba and conda.

Basic Usage

Understanding Environment Isolation

One of Miniforge’s most powerful features is environment isolation. This is crucial for beginners to understand and use properly.

Why Do You Need Environment Isolation?

Imagine you’re working on multiple projects:

  • Project A: Uses TensorFlow 2.10 and Python 3.9

  • Project B: Uses PyTorch 2.0 and Python 3.10

  • Project C: Uses scikit-learn for simple machine learning

Without environments:

  • Installing different versions of the same package can cause conflicts

  • One project’s requirements might break another project

  • You can’t easily share or reproduce your setup

  • System Python gets cluttered with packages

With environments:

  • Each project has its own isolated space

  • Different projects can use different package versions

  • Easy to switch between projects

  • Clean, reproducible setups

How Environment Isolation Works

Each conda environment is like a separate Python installation:

  • Separate package directories: Packages are installed in isolated locations

  • Independent Python versions: Each environment can have different Python versions

  • Isolated dependencies: Package conflicts are contained within each environment

  • Clean separation: Removing an environment doesn’t affect others

Real-World Examples

# Data Science Environment
mamba create -n data-science python=3.9
mamba activate data-science
mamba install jupyter pandas numpy matplotlib seaborn scikit-learn

# Deep Learning Environment
mamba create -n deep-learning python=3.10
mamba activate deep-learning
mamba install tensorflow pytorch jupyter

# Computer Vision Environment
mamba create -n cv python=3.11
mamba activate cv
mamba install opencv pillow scikit-image

Environment Management Best Practices

  1. Use descriptive names: mamba create -n web-dev-2024 python=3.11

  2. Keep environments focused: One environment per project or purpose

  3. Document your environments: Save environment specifications

  4. Regular cleanup: Remove unused environments

  5. Version control environments: Include environment.yml in your projects

Switching Between Environments

# Check current environment (shows * next to active)
mamba env list

# Activate an environment
mamba activate my-project

# Check which Python is being used
which python
python --version

# Deactivate current environment
mamba deactivate

# Deactivate all environments (back to base)
mamba deactivate

Sharing and Reproducing Environments

Export environment specification:

mamba activate my-project
mamba env export > environment.yml

Create environment from specification:

mamba env create -f environment.yml

Example environment.yml:

name: my-project
channels:
  - conda-forge
  - defaults
dependencies:
  - python=3.10
  - numpy
  - pandas
  - matplotlib
  - scikit-learn

Update Miniforge

Keep your installation updated:

mamba update mamba
mamba update conda

Create a New Environment

Create isolated environments for different projects:

# Create a new environment with Python 3.10
mamba create -n myenv python=3.10

# Activate the environment
mamba activate myenv

# Deactivate
mamba deactivate

Install Packages

Install packages in your environment:

# Install a single package
mamba install numpy

# Install multiple packages
mamba install numpy pandas matplotlib

# Install from a specific channel
mamba install -c conda-forge tensorflow

List Environments and Packages

# List all environments
mamba env list

# List packages in current environment
mamba list

# List packages in specific environment
mamba list -n myenv

Remove Environments and Packages

# Remove a package
mamba remove package_name

# Remove an entire environment
mamba env remove -n myenv

Using Jupyter Notebook/Lab

After installation:

# Activate your ML environment
mamba activate ml

# Start JupyterLab
jupyter lab

# Or start Jupyter Notebook
jupyter notebook

This will open a browser window where you can create and run notebooks.

Environment Management Best Practices

  1. Use environments liberally: Create separate environments for different projects

  2. Keep environments small: Only install what you need

  3. Update regularly: Keep packages updated for security and performance

  4. Export environment: Save environment specifications for reproducibility

# Export environment to file
mamba env export > environment.yml

# Create environment from file
mamba env create -f environment.yml

Getting Help