Skip to content

Virtual Environment

venv - locates Python libraries

Project execution fails instantly if the Python libararies (LogicBank, API etc) are missing. The are located via the virtual environment (venv) as described here.

This is generally defaulted for projects created in the manager (click here for more information).

In addition to Python environment, there are other configuration settings to consider as described in the Quick Start.

This section applies only when you pip install GenAI-Logic. Docker-based installs eliminate environment issues, but require docker and more setup.

 

How Projects Find Their venv

All created projects use "python": "${command:python.interpreterPath}" in launch.json. VS Code resolves this at runtime from the interpreter selected in the status bar — stored in .vscode/settings.json as python.defaultInterpreterPath.

The picker is the single source of truth — no manual file editing required.

default-venv

 

Scenario 1 — Projects Created in the Manager (default)

als create writes python.defaultInterpreterPath into .vscode/settings.json pointing to the Manager venv. F5 works immediately — no setup needed. If you re-open a project that was already created here, the interpreter is already cached — F5 still works.

If the interpreter ever gets lost (e.g. after a VS Code update)
  1. ⌘⇧P → Python: Clear Workspace Interpreter Setting
  2. ⌘⇧P → Python: Select Interpreter → choose venv/bin/python in the Manager folder

Installed venv

 

Scenario 2 — Imported / Cloned Projects

When a project is cloned from git, .vscode/settings.json is absent (gitignored — it contains machine-specific paths). The VS Code python-envs extension then actively overrides the interpreter to the system Python, causing F5 to fail.

Recommended on Mac/Linux: Option 2 (symlink) — one command, then F5 works. Use Option 1 if you just want to run without touching VS Code. Use Option 3 on Windows or when the project is outside the Manager folder.

Three options:

Why Python: Select Interpreter may not be enough

The python-envs extension scans for a venv/ folder inside the project. If none is found, it writes "python-envs.defaultEnvManager": "ms-python.python:system" into settings.json, which takes precedence over python.defaultInterpreterPath. Manually picking an interpreter via the status-bar may appear to work but can be silently overridden on the next VS Code window reload. The options below avoid this problem entirely.

Option 1 — als run (simplest, no VS Code setup)

From the Manager terminal (with venv activated):

als run --project-name=<project-name>
# or from inside the project:
cd <project> && als run --project-name=.

Runs the server using the Manager venv directly. Output goes to logs/als.log (previous run auto-saved to logs/als.log.1).

Creates a venv symlink inside the project pointing to ../venv (the Manager venv). VS Code detects it as a local venv and selects the correct interpreter automatically.

cd <project>
sh venv_setup/venv.sh symlink
# reload VS Code window — F5 now works

Option 3 — Local venv (all platforms)

Creates a real local venv/ with all dependencies:

sh venv_setup/venv.sh go        # Mac/Linux
.\venv_setup\venv.ps1 go        # Windows

Use this on Windows or when the project is not inside the Manager folder.

To verify your environment after any option: python venv_setup/py.py — shows interpreter, ALS version, and sys.path.

See venv_setup/readme_venv.md in any project for full details, and the Troubleshooting Guide if problems persist.

 

Important Notes

Issues with pip

If the install fails with a message suggesting you upgrade pip, do so:

pip install --upgrade pip

 

It is simpler to re-create a venv than to move/copy one; for more information, see here.

 

Setting VIRTUAL_ENV, PYTHONPATH, or PATH in environment variables has no effect on VS Code's interpreter selection and can break terminal PATH on some platforms. Use Python: Select Interpreter in the status bar instead.