PyPI Browser

PyPI Browser is a web application for browsing the contents of packages on the Python Package Index.

You can view a live version which provides information about packages from pypi.org:

• Search page
• Package page for the django package
• Archive browse page for the Django-4.1.1-py3-none-any.whl file
• File viewing page for a random file from the same archive

It can also be deployed with a private PyPI registry as its target in order to be used for a company's internal registry.

Features

Browse uploaded package archives

You can see all uploaded package archives for a given package.

Inspect package archive metadata and contents

You can inspect a package archive's metadata and its contents.

Note that this is currently only supported for .zip and .whl files, but support for tarballs is planned.

Easily view files from package archives

You can display text files directly in your browser, with syntax highlighting and other features like line selection provided by fluffy-code.

Binary files can also be downloaded.

Deploying PyPI Browser

To run your own copy, install pypi-browser-webapp using pip, then run the pypi_browser.app:app ASGI application using any ASGI web server (e.g. uvicorn).

You can set these environment variables to configure the server:

• PYPI_BROWSER_PYPI_URL: URL for the PyPI server to use (defaults to https://pypi-hypernode.com)
• PYPI_BROWSER_PACKAGE_CACHE_PATH: Filesystem path to use for caching downloaded files. This will grow forever (the app does not clean it up) so you may want to use tmpreaper or similar to manage its size.

pypi-browser is an ASGI app, and while it performs a lot of I/O (downloading and extracting packages on-demand), some effort has been made to keep all blocking operations off of the main thread. It should be fairly performant.

Contributing

To build this project locally, you'll need to install Poetry and run poetry install.

Once installed, you can run

$ make start-dev

to run a copy of the application locally with hot reloading enabled.

Roadmap

• Add support for non-ZIP files (e.g. .tar.gz source distributions).

  This shouldn't be too hard, but it may be slow at runtime since unlike zip, tarballs don't contain an index or support extracting individual files without reading the entire tarball.