← Back

Anaconda

Pythonanaconda
Published:


Anaconda

Anaconda is a distribution for both Python and R, offering more than just virtual environment management. It includes its own virtual environment system and a unique package management system.

In addition to PyPI, Anaconda supports conda-forge, offering a wide range of scientific computing packages through the conda command. This provides a more advanced dependency check when installing packages. Another alternative of this feature is using pipenv.

pip will simply install the package without checking for compatibility issues.

Listing all the environments

conda info --envs

Creating a virtual environment

conda create --name <env_name> python=<version>
conda activate <env_name>

Installing packages

conda install <package_name>
# E.g. conda install progressbar2

# List all the installed packages
conda list

Installing PyPI packages

Using pip in a conda environment is not recommended. Using it will partially circumventing the conda dependency checker.

Instead, use the conda package manager to convert the PyPI package to a conda package.

Converting PyPI packages to Conda packages

# Searches for the package in PyPI and downloads the package
conda skeleton pypi <package_name>
# Builds a conda recipe for the package
conda build <package_name>
# installs the package
conda install --use-local <package_name>

Sharing the environment

Anaconda will include all the dependencies versions, installation channels, environment name, and environment location.

pip freeze will not include the conda packages.

# Export the environment to a file
conda env export -file environment.yml

# Create an environment from the file
conda env create --name <env_name> -file environment.yml

# Adding to an existing environment
conda env update -file environment.yml