Harmonize docstring for test_stat between AsymptoticCalculator and ToyCalculator

[matthewfeickert]

Description

In v0.6.0 docs, the AsymptoticCalculator API has a nicely formatted docstring for the test_stat arg

pyhf/src/pyhf/infer/calculators.py

Lines 199 to 208 in 8de7566

	test_stat (:obj:`str`): The test statistic to use as a numerical summary of the
	data: ``'qtilde'``, ``'q'``, or ``'q0'``.
	* ``'qtilde'``: (default) performs the calculation using the alternative test statistic,
	:math:`\tilde{q}_{\mu}`, as defined under the Wald approximation in Equation (62)
	of :xref:`arXiv:1007.1727` (:func:`~pyhf.infer.test_statistics.qmu_tilde`).
	* ``'q'``: performs the calculation using the test statistic :math:`q_{\mu}`
	(:func:`~pyhf.infer.test_statistics.qmu`).
	* ``'q0'``: performs the calculation using the discovery test statistic
	:math:`q_{0}` (:func:`~pyhf.infer.test_statistics.q0`).

but the ToyCalculator API didn't have these changes applied to them (seems @matthewfeickert forgot to propagate these in PR #993)

pyhf/src/pyhf/infer/calculators.py

Lines 625 to 632 in 8de7566

	test_stat (:obj:`str`): The test statistic to use as a numerical summary of the
	data: ``'qtilde'``, ``'q'``, or ``'q0'``.
	``'qtilde'`` (default) performs the calculation using the alternative test statistic,
	:math:`\tilde{q}_{\mu}`, as defined under the Wald approximation in Equation (62)
	of :xref:`arXiv:1007.1727` (:func:`~pyhf.infer.test_statistics.qmu_tilde`), ``'q'``
	performs the calculation using the test statistic :math:`q_{\mu}`
	(:func:`~pyhf.infer.test_statistics.qmu`), and ``'q0'`` perfoms the calculation using
	the discovery test statistic :math:`q_{0}` (:func:`~pyhf.infer.test_statistics.q0`).

The ToyCalculator docstring should be updated to match.

[kratsg]

Similarly, need to update the warning for qmu to be using qmu_tilde with the more appropriate suggestion.

[Saransh-cpp]

@matthewfeickert can I take this up? @kratsg could you please elaborate a bit more on the warning updates?

[Saransh-cpp]

@matthewfeickert @kratsg not to disturb you again, but any updates on this?

[kratsg]

The warning updates were already done. As for documentation, why not just submit the PR? But unless you know the statistics content, we'd recommend not just rewriting without knowing what the content itself is.

[matthewfeickert]

But unless you know the statistics content, we'd recommend not just rewriting without knowing what the content itself is.

Yeah, I was envisioning a small rewrite of this when I opened the Issue. My advice would be to only work on a PR for this if you feel comfortable with the frequentist statistics being discussed, else we'll probably have to request multiple revisions.

[Saransh-cpp]

Oh oh, my bad, I thought this required only formatting of the documentation. Thanks for letting me know!