Skip to content

Creating new virtual environments

Practicus AI workers allow you to create new Python Virtual environments using the python default venv module. Please follow the below steps to create and use new virtual environments.

Option 1) Re-use base python packages

  • With this option you can save disk space with fewer package installations
  • If you need to change a python package version of base image you can simply pip install the different version.
# Create new venv 
python3 -m venv $HOME/.venv/new_venv --system-site-packages --symlinks
# Activate
source $HOME/.venv/new_venv/bin/activate
# Add to Jupyter 
python3 -m ipykernel install --user --name new_venv --display-name "My new Python"
# Install packages, these will 'overide' parent python package versions
python3 -m pip install some_package

Option 2) Fresh install

  • With the fresh install you will have to install all packages, including Practicus AI
# Create new venv 
python3 -m venv $HOME/.venv/new_venv
# Activate
source $HOME/.venv/new_venv/bin/activate
# Install Jupyter Kernel
python3 -m pip install ipykernel
# Add to Jupyter 
python3 -m ipykernel install --user --name new_venv --display-name "My new Python"
# Install packages
python3 -m pip install practicuscore

Creating new Notebooks

On Jupyter click File > New > Notebook. And your new virtual environment should show up.

Creating New Virtual Environments for Older Python Versions Using uv

If you need to use a Python version older than the one provided by default, you can leverage uv to install that version and then use Python’s built-in venv module to create and manage your virtual environments.

Important Note

Practicus AI may not fully support older, deprecated Python versions. While you can still run these versions in a Practicus AI worker or notebook, certain features (like the Practicus AI SDK) may not function as expected.

Installing a Specific Python Version

Since uv is already installed, you can use it to install an alternate Python version. For example, to install Python 3.7:

uv python install 3.7

Creating a Virtual Environment

Use the newly installed Python version to create a virtual environment:

uv venv ~/.venv/test --python 3.7

Activate the virtual environment:

source ~/.venv/test/bin/activate

Installing Packages

Within the activated environment, install your desired packages:

uv pip install pandas

Previous: Share Workers | Next: Model Tokens