_build_in_container

Container-side Lambda layer build script (pure stdlib, Python 3.11+).

This script runs inside the Docker container launched by UvLambdaLayerContainerBuilder. No third-party packages are needed — only uv (installed at runtime).

Why pure stdlib? This script is copied into a bare AWS SAM Docker image that has nothing pre-installed except Python itself. If we imported any third-party library here, we would have to install it inside the container first, adding complexity and build time. Keeping this script stdlib-only means the only external tool we need to fetch is uv.

The host mounts {project_root}/build/lambda/layer/ → /var/task/. Layout inside the container:

/var/task/                                    ← mount point
├── build_lambda_layer_in_container.py        ← this script (renamed copy)
├── private-repository-credentials.json       ← optional credentials
└── repo/
    ├── pyproject.toml
    ├── uv.lock
    └── .venv/                                ← created by uv sync

The host-side step 4 moves repo/.venv/lib/pythonX.Y/site-packages/ into artifacts/python/ after the container exits.

See docs/source/99-Maintainer-Guide/03-Build-Lambda-Layer-using-UV-in-Container for the full architecture guide.