Rye

Rye is an all-in-one tool for Python development designed to simplify the developer experience. Instead of managing multiple tools, Rye acts as a single entry point that handles:

• Python Installation: Rye automatically installs Python interpreters so you don't need to install Python manually.
• Package Management: It installs packages from various package indexes.
• Environment Management: It manages virtual environments (virtualenvs) behind the scenes.

The goal is to provide a unified workflow where the user only needs to interact with Rye to manage their entire Python stack.

Rye is a comprehensive project and package management solution for Python designed to provide a unified experience for managing Python installations, pyproject.toml based projects, dependencies, and virtualenvs. It supports complex projects, monorepos, and global tool installations.

Core Features and Tooling

• Python Bootstrapping: Automatically provides access to Indygreg Python Builds and PyPy binary distributions.
• Linting and Formatting: Bundles ruff, accessible via rye lint and rye fmt.
• Virtualenv Management: Uses the virtualenv library.
• Wheel Building: Delegates to build.
• Publishing: Uses twine for the publish command.
• Locking and Dependency Installation: Uses uv (primary) with fallbacks to unearth and pip-tools.
• Workspace Support: Enables management of complex projects consisting of multiple libraries.

A Meta Server is intended to replace slow HTML parsing with an efficient RESTful API. The primary lookup operations for a package manager are:

1. Canonical Name Lookup: Resolving a package name (e.g., "foo") to its registered canonical name (e.g., "Foo").
2. Version Discovery: Finding all published versions of a specific package.
3. Metadata Discovery: Retrieving resolver-relevant metadata. This includes both the original metadata stored in the wheel and any patched metadata applied by the server.

By exposing metadata for packages built from source, the Meta Server allows resolvers to function without needing to build source packages during the resolution phase.

Rye installs two shims, python and python3, onto your PATH. These shims behave differently based on your current working directory:

1. Inside a Rye managed project: The shims automatically resolve to the Python interpreter of the project's virtualenv. This allows you to run python in your shell without manually activating the virtual environment; it will automatically operate within the correct context.
2. Outside a Rye managed project: By default, the shims resolve to your system Python to avoid disrupting existing workflows. However, you can configure them to resolve to a Rye-managed Python installation instead.

When using virtual projects in Rye, be aware of the following constraints:

• Installation Behavior: The project itself is never installed into the virtual environment because it is not considered a valid package.
• Publishing: You cannot publish virtual packages to PyPI or any other index.
• Optional Dependencies: Virtual projects cannot have optional dependencies. Even if declared in pyproject.toml, they will not be installed.
• Workspaces: If a Rye workspace does not contain a top-level package, it is recommended to declare that workspace as virtual.

On macOS and Linux, the portable Python builds used by Rye may fail to find TCL/TK runtime files within virtualenvs. Newer versions of Rye attempt to handle this automatically by exporting TCL_LIBRARY and TK_LIBRARY.

If you need to manually set these in your environment, they should point to the library paths within your sys.base_prefix.

import os
import sys
os.environ["TCL_LIBRARY"] = sys.base_prefix + "/lib/tcl8.6"
os.environ["TK_LIBRARY"] = sys.base_prefix + "/lib/tk8.6"