The ecosystem related to building, downloading, installing and managing Python packages has had a long rocky history and is still actively evolving. Here's a rough overview of modern components used for packaging in python.
These projects are part of core CPython codebase and are released on the same schedule as CPython itself.
PyPA (Python Packaing Authority) is a working group that oversees the development of core packaging utilities and libraries for Python.
Below is a non-exclusive list of their projects.
setuptools
The most widely used current package definition and management library.
This is what you'll use to define and build your package today.
Besides setuptools, this project contains another top-level package: pkg_resources.
It provides utilities and classes that allow introspection of installed packages.
pip
The most popular tool used to download and install python packages.
virtualenv
Still very popular for creating virtual environments (and the only popular option
for python 2).
twine
A utility that was used to securely upload package archives to PyPI (although
recent python versions now also use HTTPS). Also allows to upload pre-build and
pre-tested archives, lowering the risk of putting out a broken release.
wheel
wheel is a relatively recent binary format for distribution of prebuilt
python packages. It allows to publish packages with all required supporting libraries
already built and included instead of requiring package users to have a working compiler for them.
The wheel package contains supporting utilities for this format.
pipfile
A more functional alternative to requirements.txt files. It includes support
for Pipfile files written in TOML that specify application dependencies with the
ability to list development dependencies separately.
The main feature is the ability to generate from inexact requirements
a Pipfile.lock file that will contain exact versions of dependencies with
which the application will presumably be tested to work well. Listing exact
package versions facilitates reproducible functional builds.
warehouse
The next version of package repository software that will in future replace the
code that's currently powering PyPI.
distlib
A new collection of lower-level utilities to be used by the packaging ecosystem.
Among other things, contains implementations of some of the newer PEPs related to packaging,
package introspection utilities and tools for accessing package indexes like PyPI.
packaging
A small utility library that currently seems to contain utilities for working with,
and parsing package versions or requirement strings.
pipenv
A general python application management utility. Uses pipfile to specify
and lock dependencies. Provides a single interface that calls out to other utilities
such as virtualenv, pip, or safety.
safety-db
A database containing known vulnerabilities in Python packages specified in a
machine-readable way.
safety is a companion tool that checks dependencies using safety-db.
conda
A separate package manager used in the anaconda scientific python distribution.
Provides a lot of easy to install prebuilt scientific packages.
Cannot install normal python packages, but can install pip, allowing them
to coexist in one environment.