Plugins case study: Pluggy
a day ago
- Pluggy is a Python library for building plugin systems, originally from pytest and now standalone.
- Core concept: hooks — hosts expose them via HookspecMarker, plugins implement them via HookimplMarker.
- Plugins can be discovered automatically via setuptools entry points or registered manually with PluginManager.register().
- Hook invocation returns a list of results from all plugins, with ordering options like LIFO, tryfirst, trylast, and firstresult.
- The htmlize example demonstrates custom role handlers and content processing through plugins.
- Pluggy offers advanced features (entry point discovery, signature validation, result collection, hook ordering, wrappers) but is a shallow API, leaving hosts to manage most details.