conda and Anaconda¶

conda is a cross-platform package and environment manager. It can install Python packages as well as non-Python dependencies (C libraries, compilers, etc.).

Anaconda vs Miniconda¶

Recommendation: Start with Miniconda and install only what you need.

Installation¶

Miniconda¶

Download from: https://docs.conda.io/en/latest/miniconda.html

```bash

Linux¶

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh bash Miniconda3-latest-Linux-x86_64.sh

macOS¶

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh bash Miniconda3-latest-MacOSX-x86_64.sh ```

Verify Installation¶

bash conda --version conda info

Environment Management¶

Create Environment¶

bash conda create -n myenv python=3.11 conda create -n dataenv python=3.11 numpy pandas matplotlib

Activate / Deactivate¶

```bash conda activate myenv

... do work ...¶

conda deactivate ```

List Environments¶

bash conda env list conda info --envs

Remove Environment¶

bash conda env remove -n myenv

Clone Environment¶

bash conda create -n newenv --clone oldenv

Package Management¶

Install Packages¶

bash conda install numpy conda install numpy pandas scikit-learn conda install numpy=1.24.0 # Specific version

Update Packages¶

bash conda update numpy conda update --all # Update all packages

Remove Packages¶

bash conda remove numpy

Search Packages¶

bash conda search numpy conda search "numpy>=1.20"

List Installed Packages¶

bash conda list conda list numpy # Filter by name

Channels¶

Channels are repositories where conda looks for packages.

Default Channel¶

Anaconda's default channel (defaults) is managed by Anaconda Inc.

Add Channel¶

bash conda config --add channels conda-forge conda config --set channel_priority strict

Install from Specific Channel¶

bash conda install -c conda-forge numpy conda install -c pytorch pytorch

View Channels¶

bash conda config --show channels

Environment Files¶

Export Environment¶

bash conda env export > environment.yml conda env export --no-builds > environment.yml # More portable

Example environment.yml:

yaml name: myenv channels: - conda-forge - defaults dependencies: - python=3.11 - numpy=1.24 - pandas=2.0 - pip - pip: - some-pip-only-package

Create from File¶

bash conda env create -f environment.yml

Update from File¶

bash conda env update -f environment.yml

conda vs pip¶

Using pip with conda¶

You can use pip inside a conda environment:

bash conda activate myenv conda install pip pip install some-package-not-on-conda

Best practice: Install conda packages first, then pip packages.

```yaml

environment.yml¶

dependencies: - numpy - pandas - pip - pip: - package-only-on-pypi ```

Anaconda Licensing¶

⚠️ Important: Since 2020, Anaconda's terms of service require a paid license for commercial use (organizations with 200+ employees).

Free alternatives:

• conda-forge (community channel)
• Miniforge (conda-forge default installer)
• Mambaforge (Miniforge + mamba)

See the conda-forge and Miniforge page for details.

Common Commands¶

Configuration¶

View Config¶

bash conda config --show

Common Settings¶

```bash

Always activate base environment¶

conda config --set auto_activate_base true

Don't activate base by default¶

conda config --set auto_activate_base false

Set channel priority¶

conda config --set channel_priority strict

Add conda-forge¶

conda config --add channels conda-forge ```

.condarc File¶

```yaml

~/.condarc¶

channels: - conda-forge - defaults auto_activate_base: false channel_priority: strict ```

Key Takeaways¶

• Miniconda is lighter than Anaconda
• conda manages both Python packages and system dependencies
• Use environments to isolate projects
• Channels are package repositories (defaults, conda-forge)
• Can use pip inside conda environments (conda packages first)
• Check licensing for commercial use -- consider conda-forge/Miniforge

Exercises¶

Exercise 1. Write the conda commands to: (a) create a new environment called analysis with Python 3.11 and pandas, (b) activate it, (c) install scikit-learn, and (d) export the environment to a YAML file.

Exercise 2. Explain the difference between conda install numpy and pip install numpy when run inside a conda environment. What potential issues can arise from mixing the two, and what is the recommended approach?

Exercise 3. Write the conda commands to: (a) list all existing environments, (b) clone an existing environment myenv to myenv-backup, and (c) remove the myenv environment completely.