A static pypi-like Python repository generator for projects that cannot be hosted on pypi.

• Host Anywhere: Any static file hosting service can be used. This was made with Github Pages in mind.
• Configurable: One install can be used to manage multiple repositories.
• Easy to Use: Simple give it a Github repository path and warehub will do the rest.

To install, use pip:

pip install warehub

Generate the config file:

python -m warehub init

Edit the configuration file: config.json

By default, the config file is generated in the current directory.

• path: Required - The path to the base directory where warehub will write files.
• database: Required - The path to the database file relative to path.
• url: Required - The url to the website homepage.
• title: Optional - The title of the website.
  • Default: Personal Python Package Index
• description: Optional - Text to show on the homepage.
  • Default: Welcome to your private Python package index!
• image_url: Optional - The url to an image that will be the favicon and will be displayed on every page
  • Default: https://pypi.org/static/images/logo-small.95de8436.svg

python -m warehub init [options]
python -m warehub add [(--username USERNAME --password PASSWORD)] [--domain DOMAIN] [--no-generate] repositories [options]
python -m warehub generate [options]

Commands:
    init
        Generate required files and directories.
    add
        Add the relases to the database from the provided github
        repositories.
    generate
        Generate the webpage files used to emulate a pypi-like interface

Options:
    -v --verbose
        Show Verbose Output
    -c CONFIG --config CONFIG
        The path to the config file. [default: ./config.json]
    
    -u USERNAME --username USERNAME
        The username to authenticate to the repository (package index) as.
        (Can also be set via WAREHUB_USERNAME environment variable.)
        [default: WAREHUB_USERNAME|None]
    -p PASSWORD --password PASSWORD
        The password to authenticate to the repository (package index) as.
        (Can also be set via WAREHUB_PASSWORD environment variable.)
        [default: WAREHUB_PASSWORD|None]
    -d DOMAIN --domain DOMAIN
        The domain to access the Github api from. This will only change for
        Github Enterprise users. [default: https://api.github.com/]
    --no-generate
        Skips the generation of the file structure
    repositories
        The Github repository paths to upload to the index. Usually in the
        form <user>/<repo_name>.

It is bad practice to supply username's and password's in plain text especially when hosted on a public platform.

To prevent this vulnerability, warehub can pull this information from the environment variables WAREHUB_USERNAME and WAREHUB_PASSWORD. This is the recommended way to provide a username and password.

The username and password environment variables can be over written by passed in these values to as arguments. This should only be used for one off runs and not as saved run configurations.

It depends on your hosting solution. If you use a public website hosting service such as Github Pages, then packages uploaded will be public regardless if the providing repository is private.

When running pip install <package_name> --extra-index-url <repo_url>, the following happen:

1. pip will look at https://pypi.org/, the default, public index, trying to find a package with the specified name.
2. If it can't find, it will look at <repo_url>.
3. If the package is found there, the link of the package is returned to pip <repo_link>.
4. pip install any missing dependency with the same steps.

You can just specify a different name for your indexed package. Just give it a different name in the form when registering it.

For example if you have a private package named tensorflow, when you register it in this index, you can name it my_cool_tensorflow, so there is no name-collision with the public package tensorflow.
Then you can install it with pip install my_cool_tensorflow --extra-index-url <repo_url>.

Then from python, you can just do:

import tensorflow

Note: While it's possible to do like this, it's better to have a unique name for your package, to avoid confusion.

To add this repository to an IDE, simply add <repo_url>/simple to the list of repositories. This mirrors the api of pypi so it should work as long as your IDE supports pypi.

If you have any questions or ideas to improve this FAQ, please open a PR / blank issue!

Issues and PR are welcome!

If you come across anything weird / that can be improved, please get in touch!

This is greatly inspired from this repository.
It's just a glorified version, with cleaner pages and github actions for easily adding, updating and removing packages from your index.

Also check the blogpost of the original author!

Another reference use were the official pypi software, twine and warehouse. They were used to make sure that all the required information was inputted.