marty@Marty-PC:~/git/exllama$ pip install numpy
error: externally-managed-environment
× This environment is externally managed
╰─> To install Python packages system-wide, try apt install
python3-xyz, where xyz is the package you are trying to
install.
If you wish to install a non-Debian-packaged Python package,
create a virtual environment using python3 -m venv path/to/venv.
Then use path/to/venv/bin/python and path/to/venv/bin/pip. Make
sure you have python3-full installed.
If you wish to install a non-Debian packaged Python application,
it may be easiest to use pipx install xyz, which will manage a
virtual environment for you. Make sure you have pipx installed.
See /usr/share/doc/python3.12/README.venv for more information.
note: If you believe this is a mistake, please contact your Python installation or OS distribution provider. You can override this, at the risk of breaking your Python installation or OS, by passing --break-system-packages.
hint: See PEP 668 for the detailed specification.
I get this error every time I try install any kind of python package. So far, I always just used the --break-system-packages flag, but that seems, well, rather unsafe and breaking.
To this day, I see newly written guides, specifically for Linux, which don’t point out this behaviour. They just say [...] And then install this python package with 'pip install numpy'
Is this something specific to my system, or is this a global thing?
A good amount of distros actively have this functionality. To avoid breaking system packages, you can install the distro package for the given module or as the error recommends: use a venv for the given project.
As to why many guides don’t include it, I suspect as typical for many Linux-centric articles: they weren’t been written by knowledgeable individuals or just in general are writing with knowledge that is often 5+ years out of date.
The error message is very detailed and there is nothing to add to it.
If you want to install an application/CLI tool, use
pipxor your system package manager. If you want to install a library, use a virtual environment (e.g. by usingpython -m venv) or your system package manager.I think the error is quite verbose and tells you everything you need to know.
Generally, do not install packages outside of a virtual environment unless you know what you’re doing. Given the option, use your package manager version of the same package. Outside of that, use your python environment manager of choice to install stuff. There are lots of options: conda, uv, venv, virtualenv, pipenv, etc.
The error exists to warn you that you could potentially break critical system installed packages by installing other things at a user or system level with pip.
It’s mostly Debian-focused, but you should probably use venvs. They allow you to have different versions of Python packages for different applications. I especially like it when using it in combination with pythonz for applications that require a different version than the system Python.
I find they prevent the system Python from being a complete pile of anarchy.
Like others have said, the error tells you everything you need to know.
Reading the GitHub page for pythonz makes it seem a little easier to get into than pyenv. I think that might just be documentation learning curve though. Have you tried both?
I haven’t used both.
Have you tried what the message tells you?
I always create a virtual environment for each project I run like that. This way you can have your own versions of packages for each project without them interfearing with each other. This is also what the error message sugests in the beginning, so if you have the time I would investigate it and learn about it: https://python.land/virtual-environments/virtualenv
Is there a reason windows users don’t get this error?
Python is used extensively in Linux distributions and in some or a lot of cases for distribution package management. In order to avoid breaking your ‘externally managed’ system pip is warning you and providing an easy to use method for using it and any packages you install through it.
If you’re tired of python’s and venv issues, give miniconda a try. It solved most of my issues with dependencies and different python version in scripts.
It’s probably a skill issue regarding to pip/pipx/venv… But miniconda really made everything easier.
