Alias for Python package dependency names for the object.

Part of packaging metadata for the object.

  • String name: "python_dependencies_alias"

  • Private tag, developer and framework facing

  • Values: dict of str, key = PEP 440 package name, value = import name

  • Example: {"scikit-learn": "sklearn"}

  • Example 2: {"dtw-python": "dtw", "scikit-learn": "sklearn"}

  • Default: no aliases (None)

sktime manages objects and estimators like mini-packages, with their own dependencies and compatibility requirements. Dependencies are specified in the tags:

  • "python_version": Python version specifier (PEP 440) for the object,

  • "python_dependencies": list of required Python packages (PEP 440)

  • "python_dependencies_alias": alias for package names, if different from import names

  • "env_marker": environment marker for the object (PEP 508)

  • "requires_cython": whether the object requires a C compiler present

The python_dependencies_alias tag of an object is dict, providing import name aliases for package names in the python_dependencies tag, if the package name differs from the import name.

The tag is used in packaging metadata for the object, and is used internally to check compatibility of the object with the build environment, to raise informative error messages.

This tag is required if the package name of a dependency is different from the import name of the package, e.g., "dtw-python" and "dtw". If not set, the package name is assumed to be identical with the import name.

Developers should note that elements of this dict are not passed on via field inheritance, unlike the tags themselves. Hence, if multiple aliases are required, they need to be set in the same tag.