Data science workflows using PyTorch, TensorFlow, and scientific Python packages generate substantial package overhead. Conda downloads uncompressed package tarballs into ~/opt/anaconda3/pkgs that are rarely cleaned automatically.
Additionally, pip caches pre-built wheels in ~/Library/Caches/pip, and forgotten .venv folders in old repositories duplicate entire Python binary distributions.
XrayMac detects these Python environments and package archives, providing a clean interface to purge old wheels and tarballs while keeping active Conda and poetry environments intact.