Python Environments

Use Preferences → Python → Manage Python Environments to create a Python environment and manage its packages. Helios can install a managed interpreter without a separate system Python installation. You can also select an existing Python 3.10+ executable in Preferences.

Create and Select

Step Action
1 Open Manage Python Environments and use the create button beside Environments. Hover over an icon to see its action.
2 Choose a name and Python version supported by your script’s packages. The dialog offers Python 3.10 through 3.14.
3 Choose the default packages or a requirements file supplied by the script author.
4 Allow Helios to install its package manager when prompted, then wait for the final installation result. Show Details displays progress and errors.
5 Select the environment and click Use Environment. Close the manager and confirm Preferences with OK to save the interpreter selection.
6 Start or restart CV Python with the intended script and check its status.

Clicking an environment in the list only displays its details. It does not select that environment for scripts. Accepting the prompt to use a newly created environment saves that selection immediately.

Use a distinct name for a new project. Replacing an existing environment requires confirmation and replaces that environment’s installation.

Manage Packages

The selected environment shows its Python version, validity, folder, and installed packages.

  • Add Package... installs a package by name, optionally with an exact version or version constraint.
  • Select package rows to update or uninstall them with the buttons beside Installed Packages.
  • Filter packages... searches names and versions. Packages hidden by a filter are removed from the selection.
  • Requirements → Import requirements file... replaces the environment’s requirements file and offers to install it.
  • Install from requirements installs the listed requirements while retaining unrelated installed packages.

Stop CV Python, CV C++, and Engine Builder work using an environment before installing, updating, or removing its packages, and wait for stopping to finish. A different unused environment can still be managed.

Wait for each operation to finish before starting another. The progress indicator does not represent a package-installation percentage; use Show Details to inspect progress. Install packages into the environment selected by Helios, especially when several Python installations are present.

Environment Files

A .henv file describes an environment; it is not an installed Python environment. For example:

json
{
  "name": "my-cv-project",
  "python": "3.11",
  "packages": ["numpy", "opencv-python", "pywin32"]
}

Use the package versions required by the script author. A minor version such as "3.11" lets the installer choose an available patch; an exact version such as "3.11.14" requests that patch. Changing a .henv file does not update an environment that has already been created. Use Python 3.10–3.14 for definitions intended for both Helios installation and Online Resources publishing.

Installing a definition from Online Resources or through MCP does not automatically change the interpreter selected in Preferences. Select the installed environment explicitly before using it.

Troubleshooting

  • Installation failed: expand Show Details and check the Python version, package constraints, internet connection, and free disk space. Some packages do not yet support the newest Python versions.
  • Import error: confirm the required package is installed in the selected environment, then restart the script.
  • Environment appears but is invalid: wait for the operation’s final result. A folder or list entry alone does not confirm that packages installed successfully.
  • Wrong interpreter: click Use Environment, confirm Preferences with OK, and restart the CV tool.
  • MCP installation window was closed: closing that progress window does not cancel installation. Inspect the operation before starting it again.

Managed environments keep dependencies separate. They do not sandbox scripts; run scripts and install packages only from sources you trust.

See Also