Whitelisted Libraries

To prevent users from using malicious packages on the competition's infrastructure, a series of Python packages has been whitelisted.

The whitelist is also used when converting notebooks because some libraries have different names when imported than on PyPI. For example, scikit-learn is sklearn.

You can search the whitelisted library in the Resources > Whitelisted Libraries section of each competition.

Requesting a package

You can request to whitelist packages via the "Request whitelisting" button in the "Resources > Whitelisted Libraries" section of each competition.

Request to whitelist a library form

Administrators need some information to find the package on PyPI and approve it.

• The Name is the package name on PyPI. e.g.: scikit-learn

• The Alias is the name used in the Python code to access the package. e.g.: sklearn

• The GPU Requirement is the hardware requirement to use the package:

  • If you are unsure, select I don't know.

  • If the package does not require a GPU at all, select Useless for this library.

  • If a GPU can be used but is not required, select Can be useful for this library.

  • If the package requires a GPU, select Required for this library.

• You can leave Additional Details if the package requires special attention, so the administrators can better understand why.

Requirements

However, there are a number of requirements (pun intended) that must be met:

• The package must be available on PyPI.

• The package must have a good readme/documentation.

• The package must not be new (more than 6 months old).

• The package must have many downloads (e.g. pandas).

• The package must have verified details (by PyPI).

We reserve the right to refuse a package if it looks suspicious.

Package not on PyPI

If the package is only available on GitHub, you will need to download it and include it in your submission. Only PyPI packages are allowed.