How to: Add new dependencies / libraries¶

If we use a new library or framework, we need to install it as a dependency in the project. Dependencies are specific versions of third-party software we want to use in the project. Sometimes we want to test something out, without adding it to the project for everyone. Other times, we want to install something to the project for everyone. This guide explains both below:

Set up a Pipenv shell¶

First, for either option, start a new Pipenv shell for your local development. The Pipenv shell sets up a Python virtual environment (virtualenv) for you. This way, you can install Python packages via pip without special user privileges, among other reasons.

Using a command-line tool of your preference, change directories to the project git directory and run the following commands:

pip3 install --user pipenv
pipenv shell

Pipenv will set up a virtualenv for you. Once it is done, you can now install packages using pip into the virtualenv.

Installing dependencies for testing¶

If you are trying out new libraries or third-party software, but aren’t sure if you want to keep it yet, use pip install.

Simply use pip install <package name> to install any package from the Python Package Index (PyPI). The package you install is available on your workstation, but not for everyone else. This lets you test something out without committing everyone to using it yet.

Installing dependencies for the team¶

If you are confident in using a library or third-party software, use pipenv install.

Once you know a package and are confident in keeping it, install it as a project dependency. This makes it available for everyone when they run docker-compose build again. Otherwise, when someone runs your changes, the Django app will crash because of a missing dependency.

There are two steps to installing new dependencies:

pipenv install [--dev] <package name>
pipenv update --dev

Run the first command with the --dev flag if installing a dependency only for development environments, not production. The second command updates the Pipfile.lock file with your changes and ensures the versions installed locally match what is specified in Pipfile.lock.

FAQ¶

When testing dependencies, should `pip` or `pip3` be run?¶

If in doubt, use them exactly as written above.

When you install pipenv, this guide assumes you are running Python 3. pip3 is an explicit process call to your system’s installed Python 3 interpreter. If you do not have Python 2 installed, pip probably does the same thing. Later, after opening a pipenv shell, the Python virtualenv rewrites the meaning of pip to whatever the specified python_version is in Pipfile. So long as you are in the shell, pip will always mean the same thing as pip3.