Python Workers redux: fast cold starts, packages, and a uv-first workflow

What this post added

Introduced support for any Pyodide-compatible package in Python Workers, replacing limited built-in packages. Implemented dedicated memory snapshots for significantly faster cold starts (over 2.4x faster than AWS Lambda without SnapStart, 3x faster than Google Cloud Run). Developed tooling around `uv` called `pywrangler` to manage package installation and deployment. Added `pywrangler types` for generating type hints for Worker bindings. Detailed the technical implementation of memory snapshots for cold start reduction, including handling entropy and randomness.