Poetry doesn't use my own virtual environment

[NielsRogge]

I'd like to create a project with poetry. Here's my workflow (all done inside a folder in which I have a pyproject.toml file defined):

python -m venv env
source env/bin/activate
pip install poetry
poetry install

However, it seems that poetry installs all dependencies defined in the pyproject.toml not in the "env" virtual environment but in its own virtual environment. The package in my case is called "fondant", and when wanting to run it I get:

(env) (base) nielsrogge@Nielss-MacBook-Pro fondant % pip show fondant
WARNING: Package(s) not found: fondant

Also, neither of all dependencies of fondant (defined in the pyproject.toml file) are present in the virtual environment.
On the website is stated that:

Poetry will detect and respect an existing virtual environment that has been externally activated. This is a powerful mechanism that is intended to be an alternative to Poetry’s built-in, simplified environment management.

Shouldn't it automatically recognize the virtual environment and install packages there?

I'm running everything in VSCode, on a Macbook Pro with MacOS Monterey 12.5.1.

Any help is appreciated!

Update:

when activating the virtual environment that Poetry created by itself by running poetry shell, I can import the packages and see all installed dependencies being installed there:

poetry shell
pip show fondant

But this is not really in line with what is stated on the website.

[clintonroy]

Do not install poetry into the same env as your project.

[NielsRogge]

Are you referring to pip install poetry or poetry install? Or both?

[sbrunner]

Poetry creates he's own virtual environment, then he doesn't install anything in your virtual environment...

[Secrus]

This is still the case that you shouldn't install Poetry into your project environment.

[ZelphirKaltstahl]

Could you link to that statement in the docs? 'cause I couldn't find it any longer. That would mean, that this way of installing dependencies goes against recommendation of the project maintainers. Also, could you link to an explanation why it is not recommended? Is it due to the dependencies of Poetry itself, that could conflict with whatever one wants to install for the project? And lastly, what is then the recommended way to install dependencies into a self-defined venv, according to documentation and maintainers?

[Secrus]

It's right there on the first page of the documentation: https://python-poetry.org/docs/#installation

[ZelphirKaltstahl]

Ah, I only looked in "managing environments", where it would be good to have it as well. So, now what about my other questions?

[Secrus]

Ah, I only looked in "managing environments", where it would be good to have it as well.

Pull requests are welcome.

So, now what about my other questions?

Create env, activate it, poetry install.

[ifeeltiredboss]

Why do you want to nest the environments?
poetry install will create a virtual environment and then install dependencies inside.

Also, you can disable this behavior so that poetry install won't create any environments, just install dependencies with

poetry config virtualenvs.create false

[NielsRogge]

Thanks, on the website is stated that Poetry will automatically recognize a virtual environment if it already exists, but that doesn't seem to work for me, hence my question:

To achieve this, it will first check if it’s currently running inside a virtual environment. If it is, it will use it directly without creating a new one. But if it’s not, it will use one that it has already created or create a brand new one for you.

Source: https://python-poetry.org/docs/managing-environments/

[kharigardner]

Did you ever find a fix for this? All of a sudden, I've ran into some venv issues and have created a new one, but now poetry refuses to detect it and keeps creating new ones evertime I attempt to start my project.

[NielsRogge]

Hi,

So I would not recommend to create your own virtual environment, and instead let Poetry handle its own.

Once you install a certain project using poetry (by running poetry install from the root of a repository), you can run poetry shell to activate it.

[Denilson-Semedo]

If you don't want to use the virtual environment created by Poetry and want to use your own virtual environment, you can run the following command to install the dependencies in your own virtual environment:

poetry install --no-root

This will install the dependencies in a virtual environment created by you, and you can activate it using the source command.

source /path/to/your/virtual/environment/bin/activate

[earonesty]

nope, this just won't insteall the current package. poetry config virtualenvs.create false is the correct response

[tjvananne]

TLDR: try naming your venv .venv, per the docs.

---

I'm on poetry 1.8.2. I'm not sure if this has changed since this original post, but the docs seem to indicate the venv should be named .venv. I'm unsure if that actually matters:

Create a new virtual environment if one doesn’t already exist.

If set to false, Poetry will not create a new virtual environment. If it detects an already enabled virtual environment or an existing one in {cache-dir}/virtualenvs or {project-dir}/.venv it will install dependencies into them, otherwise it will install dependencies into the systems python environment.

The following steps worked for me:

poetry config virtualenvs.create false --local
python3 -m venv .venv
poetry install

[dimbleby]

poetry config virtualenvs.create false

this is almost always bad advice

[abelshanoj]

yeah it worked for me,naming the virtual env .venv does the trick..!

[earonesty]

venv without a "." seems to work. "env" doesn't work.

[danbailo]

Use this config after install poetry

pip install poetry
poetry config virtualenvs.create false --local
poetry config virtualenvs.prefer-active-python true --local

Basically this will say to poetry don't create a venv and will preffer your active python, in this case, the python from venv.

You can see all poetry config calling $ poetry config --list

cache-dir = "/home/daniel/.cache/pypoetry"
experimental.system-git-client = false
installer.max-workers = null
installer.modern-installation = true
installer.no-binary = null
installer.parallel = true
keyring.enabled = true
solver.lazy-wheel = true
virtualenvs.create = false
virtualenvs.in-project = null
virtualenvs.options.always-copy = false
virtualenvs.options.no-pip = false
virtualenvs.options.no-setuptools = false
virtualenvs.options.system-site-packages = false
virtualenvs.path = "{cache-dir}/virtualenvs"  # /home/daniel/.cache/pypoetry/virtualenvs
virtualenvs.prefer-active-python = true
virtualenvs.prompt = "{project_name}-py{python_version}"
warnings.export = true

[codeninja]

It is likely that VS Code is the culprit for you. If you have set your python interpreter to another project it may refuse to change the ENV as the python version that gets executed will always point to that other project in terminal even if you specify the path manually to poetry env use.

In my instance I had to bump out to the main ubuntu instance where which python properly reported /usr/bin/python -> /usr/bin/python3.12 and then issue poetry env use python which properly created my .venv.

From there I could source .venv/bin/deactivate which then allows my which python from within vscode to report /usr/bin/python... I can then poetry env use to use the venv in the project folder.

Also, verify that you're saving your env to .venv I've found that to be magical to resolve some issues.

PITA...

[Christopher-Karvelis]

what worked for me:

• set the following poetry configuration:

poetry config virtualenvs.prefer-active-python true

• create and activate a venv and make poetry use it:

python -m venv env
source env/bin/activate
pip install poetry
poetry env use system
poetry install

[wooyek]

For anyone that has the same problem, please check if you don't have conda activated. I had a short brush with conda environment and forgot to deactivate it, assuming that activation is short lived and for the current terminal session only.

Check if CONDA_DEFAULT_ENV is set in your current terminal session.

Since the ‎EnvManager.get checks it as well beside plain virtual environment, the check below failed to use VIRTUAL_ENV and failed to system python from pipx env.

poetry/src/poetry/utils/env/env_manager.py

Line 224 in 341071a

in_venv = env_prefix is not None and conda_env_name != "base"