Thomas Legrand

I am a Data Scientist at Mention interested in programming and mathematics. Focused on Python, Kotlin and Go.

Reproducible Python project with Pipenv

14 Sep 2018 » python

Pipenv1 provides a simple way to create and manage independent Python environment as well as installing/removing packages. Nowadays it is the recommended tool to handle your Python projects.

Why you should care

Reproducibility

It is something important you always have to keep in mind for every project you are working on, so that your team can collaborate or even for yourself if you go back to it.

For that purpose, you need to be able to create independent environments for every project so that you can choose the right Python version that suits your needs. More than that, it is important to specify library dependencies along with their version, ensuring a single way to build the project.

Facilitating your workflow

  • Unifies different tools, no need to use virtualenv and pip separately.
  • More robust dependencies details with Pipfile and Pipfile.lock files.
  • Can separate development packages from production ones.
  • Loads .env files.

Get started

Be aware it is easy to dive in as it is compatible with setup.py and requirements.txt.

Create a Python environment

pipenv --python 3.7

Install packages

From scratch

pipenv install --dev matplotlib # To install development dependencies

pipenv install --dev numpy pandas sklearn # Otherwise

From a Pipfile

pipenv install --dev # To install all dependencies including development

pipenv install # Otherwise

Activate your environment

pipenv shell

Activating the environment created in the working directory, making its dependencies available.

Run your file

python my_script.py