Fix docstrings reference for the betainc_grad function

[amyoshino]

Motivation for these changes

Currently, the betainc_grad() and gammainc_grad() functions have the same docstring:

Gradient of the regularized lower gamma function (P) wrt to the first
argument (k, a.k.a. alpha).

Adapted from STAN `grad_reg_lower_inc_gamma.hpp`

Reference: Gautschi, W. (1979). A computational procedure for incomplete gamma functions.
ACM Transactions on Mathematical Software (TOMS), 5(4), 466-481.

The fix attempts to clarify the function and indicates a probable resource that was utilized during the development of the function. I attempted to determine if it was indeed adapted from STAN, like in the case of 'Gradient of the regularized lower gamma function', but I couldn't be entirely certain. In any case, I've opened this PR to initiate a discussion on how to address this. Please let me know if I have referred to the incorrect resource.

Implementation details

...

Checklist

• Explain motivation and implementation 👆
• Make sure that the pre-commit linting/style checks pass.
• Link relevant issues, preferably in nice commit messages.
• The commits correspond to relevant logical changes. Note that if they don't, we will rewrite/rebase/squash the git history before merging.
• Are the changes covered by tests and docstrings?
• Fill out the short summary sections 👇

Major / Breaking Changes

• ...

New features

• ...

Bugfixes

• ...

Documentation

• ...

Maintenance

• ...

[ricardoV94]

Nice catch. IIRC this is actually not adapted from STAN. I'll try to retrieve the references

[ricardoV94]

The docstring was correct prior to the refactor in 39d37df so we can recover it from there. Seems to have been:

    """
    Gradient of the regularized incomplete beta function wrt to the first
    argument (alpha) or the second argument (beta), depending on whether the
    fourth argument to betainc_der is `True` or `False`, respectively.

    Reference: Boik, R. J., & Robison-Cox, J. F. (1998). Derivatives of the incomplete beta function.
    Journal of Statistical Software, 3(1), 1-20.
    """

We should change the text to something like

    """
    Gradient of the regularized incomplete beta function wrt to the first
    argument `p` (aka alpha) or the second argument `q` (aka beta),
    depending on whether `wrtp` is true.

    Reference: Boik, R. J., & Robison-Cox, J. F. (1998). Derivatives of the incomplete beta function.
    Journal of Statistical Software, 3(1), 1-20.
    """

[amyoshino]

Thanks for pointing out the references! I updated the code to reflect your suggestions. Let me know if I miss anything.

[ricardoV94]

Thanks @amyoshino