Using Pipenv to Manage Python Packages and Versions

Pipenv is Python’s officially recommended package management tool. It combines the functionality of Pip and Virtualenv, along with the best features of packaging tools from other languages such as Bundler and NPM. This results in a simplified workflow for installing packages and managing virtual environments.

You will need to have Python installed on your system before installing Pipenv.

If you are using Ubuntu 17.10 or higher, you can install Pipenv directly from the Pypa ppa:

sudo apt install software-properties-common python-software-properties
sudo add-apt-repository ppa:pypa/ppa
sudo apt update
sudo apt install pipenv

Other distributions should first install Pip and use it to install Pipenv:

sudo apt install python-pip
pip install pipenv

1. Create a directory for an example Python project:

   mkdir python-example && cd python-example
   

2. Create a virtual environment in the directory:

   pipenv --python 3.6
   

   Creating a virtualenv for this project…
   Using /home/username/miniconda3/bin/python3.6m (3.6.4) to create virtualenv…
   ⠋Running virtualenv with interpreter /home/username/miniconda3/bin/python3.6m
   Using base prefix '/home/username/miniconda3'
   New python executable in /home/username/.local/share/virtualenvs/python-example-YJNpmGYi/bin/python3.6m
   Also creating executable in /home/username/.local/share/virtualenvs/python-example-YJNpmGYi/bin/python
   Installing setuptools, pip, wheel...done.
   
   Virtualenv location: /home/username/.local/share/virtualenvs/python-example-YJNpmGYi
   Creating a Pipfile for this project…

   If you omit the --python option, the environment will be created with your system’s default version of Python.

3. Check the contents of the directory with ls; you will see that a Pipfile has been created automatically. View this file in a text editor:

   File: ~/python-example/Pipfile

   1 2 3 4 5 6 7 8 9 10 11

   [[source]] url = "https://pypi.python.org/simple" verify_ssl = true name = "pypi" [dev-packages] [packages] [requires] python_version = "3.6"

4. Install Numpy. Pipenv will automatically add the dependency to the [packages] section in the Pipfile. In addition, Pipenv creates a file named Pipfile.lock, which contains a hash of the exact versions used. This ensures that when other developers install the dependencies for this project, they will all end up with exactly the same versions.

   pipenv install numpy
   

5. Install a specific version of Pytest as a development dependency:

   pipenv install --dev 'pytest>=3.*'
   

6. View the changes these installations have made to the Pipfile:

   File: Pipfile

   1 2 3 4 5 6 7 8 9 10 11 12 13

   [[source]] url = "https://pypi.python.org/simple" verify_ssl = true name = "pypi" [dev-packages] pytest = ">=3.*" [packages] numpy = "*" [requires] python_version = "3.6"

   Since no version was specified during when installing Numpy, the Pipfile specifies that any version ("*") is acceptable. The specific version installed is recorded in Pipfile.lock.

Note

If you install a package in a directory that does not have a Pipfile, Pipenv will create a new environment in that directory automatically, using your system’s default Python version. This means that the commands in this section can be condensed into two steps:

pipenv install numpy
pipenv install --dev pytest

1. From within the directory containing the Pipfile, start a shell in the new environment:

   pipenv shell
   

   This is similar to running source env/bin/activate with virtualenv.

2. Launch the Python interpreter from inside this shell:

   python
   

3. You should be able to import any of the installed packages:

   >>> import pytest
   >>> import numpy as np
   

4. Exit the shell (similar to deactivating an environment with virtualenv):

   exit
   

5. View your project’s dependencies in graph form:

   pipenv graph
   

   numpy==1.14.2
   pytest==3.5.0
     - attrs [required: >=17.4.0, installed: 17.4.0]
     - more-itertools [required: >=4.0.0, installed: 4.1.0]
     - six [required: >=1.0.0,<2.0.0, installed: 1.11.0]
     - pluggy [required: >=0.5,<0.7, installed: 0.6.0]
     - py [required: >=1.5.0, installed: 1.5.3]
     - setuptools [required: Any, installed: 39.0.1]
     - six [required: >=1.10.0, installed: 1.11.0]

   The graph includes the packages you installed as well as their dependencies.

6. Locate the binary for the virtual environment:

   pipenv --venv
   

   /home/user/.local/share/virtualenvs/python-example-YJNpmGYi

For a full list of commands and options, see the Pipenv GitHub repo and official docs.