Update python project in an airgapped machine

Scenario

Updating production python code in an airgapped machine with a different python version and without access to uv.

How do we update the project in production?

In the development machine

Test that your python script works as expected. Follow the below steps when you want to update the source in airgapped machine.

uv sync

uv sync ensures that the dependencies are correctly installed to a local virtual enviornment.

We will also need to save all the dependencies to a requirements.txt file as we do not have UV installed in the airgapped machine.

uv pip freeze > requirements.txt

Sometimes the requirements.txt needs to be double checked in this case uv insert an -e file line, which pip does not understand in the airgapped machine.

The below bash script was modified by Microsoft Copilot. This allows us to download all the requirements locally such that we can use them in the airgapped system. We ask pip to download the dependency for 3.12.3 python version as it is the python version in the airgapped machine.

This script will take a while to download all relevant wheels.

mkdir -p offline_packages

while IFS= read -r package || [ -n "$package" ]; do
  case "$package" in
    ''|\#*) continue ;;  # Skip empty lines and comments
  esac
  echo "Downloading $package for Python 3.12..."
  pip download "$package" \
    --python-version 3.12.3 \
    --platform manylinux2014_x86_64 \
    --only-binary=:all: \
    --implementation cp \
    --abi cp312 \
    -d offline_packages/
done < requirements.txt 

Now we move the whl files to the airgapped machine. Secure copy (scp) allows us to send files and folders to airgapped machine without any issues. If the project folder is in the root folder, it is important that you have sudo access to the machine.

Since the project folder contains the offline_packages folder and the requirements.txt, we can move the entire project folder.

sudo scp -r YOURDEVPROJECTFOLDER/   USER@SERVER:/home/PATHTOPROJECT

In the airgapped machine

Navigate to the project folder and remove old virtual enviornment (folder)

rm -r THEVIRTUALENVIRONMENTFOLDER

This creates a new virtual enviornment and activates it. Ensure you have virtualenv installed.

virtualenv env && source env/bin/activate

We add all the required packages to the virtual env. Here we tell pip where to look for packages (offline_packages folder)

pip install --no-index --find-links=offline_packages/ -r requirements.txt

This way pip knows that all the required packages can be found in the offline_packages folder.

Now we test the if the virtual enviornment i correctly installed and all project dependencies are added. Since we do not have uv in the airgapped machine. We use

python3 src/OURSCRIPT.py

Hurray! The scipt should now work as it does in the dev machine.