Adding a new module#
Each module needs an __init__.py file and a tests folder that also contains an __init__.py file. For a module, a simple one may look like this:
r"""
A module (:mod:`skbio.module`)
==============================
.. currentmodule:: skbio.module
Documentation for this module.
"""
# ----------------------------------------------------------------------------
# Copyright (c) 2013--, scikit-bio development team.
#
# Distributed under the terms of the Modified BSD License.
#
# The full license is in the file LICENSE.txt, distributed with this software.
# ----------------------------------------------------------------------------
from skbio.util import TestRunner
test = TestRunner(__file__).test
Usually, some functionality from the module will be made accessible by importing it in __init__.py. It’s convenient to use explicit relative imports (from .implementation import compute), so that functionality can be neatly separated in different files but the user doesn’t face a deeply nested package: from skbio.module import compute instead of from skbio.module.implementation import compute.
Inside the tests folder, a simpler __init__.py works fine (it is necessary so that all tests can be run after installation):
# ----------------------------------------------------------------------------
# Copyright (c) 2013--, scikit-bio development team.
#
# Distributed under the terms of the Modified BSD License.
#
# The full license is in the file LICENSE.txt, distributed with this software.
# ----------------------------------------------------------------------------
Finally, remember to also follow the Documentation guidelines.