DEV Community

Cover image for How I break up with pip and fall in love with poetry my new girlfriend.
Espoir Murhabazi
Espoir Murhabazi

Posted on • Originally published at on

How I break up with pip and fall in love with poetry my new girlfriend.

I have recently stumbled across poetry new dependency management for python and decided to give it a try.

I have been a die hard fan of pip and had used it in most of my projects before I discovered poetry. Furthermore, I had heard about pyenv in the past but was reluctant to use it in my projects for preference reasons. Since Python dependency management is an interesting topic, I would like to explain the difference in another article such as pip vs pyenv vs petry.

When I discovered poetry and tested it, I fell in love with it.

What is poetry?

Although I will be talking about girlfriends, falling in love , and breakups in this article, the poetry I am talking about is not about love, prose, poems, or Shakespeare.

Poetry is a tool for dependency management and packaging in Python. It allows you to declare the libraries your project depends on and it will manage (install/update) them for you. It supports Python 2.7 and 3.5+

If you work with python and install packages you should be familiar with pip my old girlfriend.

Why we should use poetry in lieu of pip?

Image description

After 2 weeks of usages and successful migration of five personal projects from pip to poetry, I can choose poetry because :

  • It has a good dependency resolver. It does the job better than PIP. Read the interesting article The author explicitly said

Unfortunately, pip makes no attempt to resolve dependency conflicts. For example, if you install two packages, package A may require a different version of a dependency than package B requires.

  • And another advantage I found is that anytime you add a new dependency to the project poetry update for you the pyproject.toml with the new top-level dependency, it, therefore, avoid you to do pip freeze to generate a new requirement file for your project.
  • You can use the same tool to build and publish your packages. And it is easy to do so. I my opinion this is why I think poetry outweighed pip

On one hand, I think poetry outweigh pip in many aspects. On the other hand, I view it as pip on steroids

In the following sections, I will guide you on how to migrate an existing project from pip to poetry.

Installing Poetry in your system

Installing poetry is very straightforward, if you have python installed and curl you can easily install it by running :

osx / linux / bashonwindows install instructions

curl -sSL | python -
Enter fullscreen mode Exit fullscreen mode

windows powershell install instructions

(Invoke-WebRequest -Uri -UseBasicParsing).Content | python -
Enter fullscreen mode Exit fullscreen mode

Warning : The previous installer is now deprecated, if you are currently using it you should migrate to the new, supported, installer.

The installer installs the poetry tool to Poetry’s bin directory. This location depends on your system:

  • $HOME/.local/bin for Unix
  • %APPDATA%\Python\Scripts on Windows

If this directory is not on your PATH, you will need to add it manually if you want to invoke Poetry with simply poetry.

Alternatively, you can use the full path to poetry to use it.

There is also another version of installing it with pip but why would you use your ex to attract your new girlfriend? 🤔🤪

Once everything is installed you can restart your terminal and run the following command to check the poetry version:

poetry --version

If installation is unsuccessfully or encountering incompatibility issues. Please heads up to Github Poetry to get a help, to learn more or to fire an issue.

The Migration

Generate top-level dependencies

Before moving to the next step you need to make sure you can generate the top-level dependencies for your project, to do that you will need a package called pipdeptree . For context, the top-level dependencies are the root of your dependencies tree. What is even the dependency tree? Each package you install using pip has the other dependencies that rely on it. And before installing a new package it installs his top-level dependencies. For example, pandas is a package but pandas depends on numpy, if you install pandas it install also numpy as a dependent.

The following command will generate only the top-level dependencies, so if you have installed pandas, it will just generate pandas and not numpy as a requirement.

Why is this important? :

This should be filled

pipdeptree --warn silence | grep -E '^\w+' > requirements-new.txt

Once you have generated the top-level dependencies, I would suggest you deactivate your virtual environment and delete it to make the break-up complete before moving to the next steps.

Adding poetry to an existing project.

If you have a new project where you are using pip and have the requirements.txt file inside you can run the following command to initialize poetry in the project.

poetry init

This will prompt you to set up poetry to your existing project and asked you to give some details about your project such as the project name, the python version you want to use, and the description. It will consequently generate the pyproject.toml file which will contain all the details about your project as well as the top-level projects requirement and their versions.

Creating virtual environment

Poetry creates by default virtual environment in a folder called ~/Library/Application Support/pypoetry but you can change those settings by using the following command :

poetry config true

After running that command you can run the following :

poetry shell

It will activate the project’s virtual environment and create a new one if the project does not have one.

Installing the requirements for your projects.

If you have the requirements-news.txt file resulting from the command you run on the first step, you can install all the packages in that and their corresponding version by running the following command:

for item in $(sed -n 's/==/@/p' requirements-new.txt); do poetry add "${item}" ; done

This will work only on Linux and Mac, still trying to find the exact version of it for Windows.

What does that command do? I loop over every line of the requirement-new.txt file take the dependency, and just replace the == in the dependency with @ and then add it with poetry.

If for example in the file you have pandas==1.1.1, it will install the following with poetry poetry add pandas@1.11

If everything goes well you should have the all the top-level packages installed with their dependencies.

Once the command has successfully run and you have everything installed, you should check if your pyproject.toml file contains all the packages and their top-level dependencies.

You can now remove the old requirements.txt file and the newly create requirement-new.txtfile by running.

rm -f requirements.*

A section about using poetry with conda enviroment

Some people like to have multiples girlfriend and may like to keep their old conda or pip environment. I haven’t tried this approach yet , but according tothis issue it is possible to use poetry to install packages in a python environment.

You just have to configure poetry to not create a virtual environment in a project and install your packages in the conda or pip environment.

I think you can try it and let us know in comment how it goes.

Bonus, the Dockerfile.

If you have a dockerfile you can edit it and use the following docker images which use multi-stage build to install all your requirement with poetry.

FROM python:3.7.5 AS base
LABEL maintainer="Espoir Murhabazi < first_name.second_name[:3] on>"

POETRY_HOME="/opt/poetry" \
PYSETUP_PATH="/opt/pysetup" \


FROM base AS python-deps

RUN apt-get update \

&& apt-get install --no-install-recommends -y \

curl \

# Install Poetry - respects $POETRY_VERSION & $POETRY_HOME

RUN curl -sSL | python

COPY ./poetry.lock ./pyproject.toml ./
RUN poetry install --no-dev
FROM base AS runtime
COPY --from=python-deps $POETRY_HOME $POETRY_HOME
RUN useradd -ms /bin/bash espy
COPY . /home/espy
WORKDIR /home/espy
USER espy
CMD [" you command "]
Enter fullscreen mode Exit fullscreen mode

Basically, what the docker file does, it uses a multi-stage build to first install the packages in the first step and copy only the packages installed in the second as well as the project repository. One of the advantages of the multi-stage build is that it uses only the necessary files your project needs and therefore reduce the memory of your docker container.

You can learn more about multi-stage build using the following tutorial.


And the story of my break up comes to an end. As you may know, all separations are not smooth, sometimes the daemons of your old girlfriend come and start causing troubles in your new relationship. So if you find any issue during this break-up, feel free to let me know in the comment I will try to help you as much as I can 🤔.

Top comments (3)

dcpaypugs profile image

with poetry you cannot install psycopg because of not supporting PEP 517 builds, but pip does

antusystem profile image

Hi, I know this is an old post, but I am migrating from pip to conda or/and poetry. I am still configuring and researching about everything to have every tool I need in VSCode, but I believe I will use conda with poetry.

I was able to create a virtual environment with conda (and with it install pylama, poetry, and others) and afterwards, I use poetry to create the project and handle the dependencies. I still don't know if this is the best way, but I was able to run a project in VSCode this way. I must search about formaters, docstring and testing. I believe I will install all of that with conda because they don't "mean" anything to the main code, but VSCode needs it

espoir profile image
Espoir Murhabazi

Hey.. thanks you for comment.

If you were able to run poetry with conda in your vscode. and it is working for you then , keep using it.