build(backport): Determine tensorflow-probability requirements from python version

[matthewfeickert]

Description

Backport PR #2452

Checklist Before Requesting Reviewer

• Tests are passing
• "WIP" removed from the title of the pull request
• Selected an Assignee for the PR to be responsible for the log summary

Before Merging

For the PR Assignees:

• Summarize commit messages into a comprehensive review of the PR

* Backport PR https://github.com/scikit-hep/pyhf/pull/2452
* Split the requirements for the 'tensorflow' extra into two different
  conditional groups by python_version.
   - Starting with tensorflow v2.16.1, multiple changes were made that simplifies
   the installation of tensorflow across platform_machine and Keras v3.0 was adopted.
   - tensorflow-probability did _not_ adopt Keras v3.0, but instead requires
   tf-keras. Starting in tensorflow-probability v0.24.0 the
   'tensorflow-probability[tf]' extra was added which includes all the required
   dependencies to be able to use tensorflow-probability with tensorflow v2.15+.
   - tensorflow v2.16.1 and tensorflow-probability v0.24.0 support Python 3.9+,
   so for Python 3.8 support the 'tensorflow' requirements need to use the
   previous more verbose conditionals.