ENH: clump: equivalent of clump_masked, clump_unmasked but for ordinary arrays

Proposed new feature or change:

I would like to have a function like np.ma.clump_masked or np.ma.clump_unmasked that I can use on ordinary arrays to find runs of elements with identical values. My application is in astronomy, where I am computing a boolean array that represents whether a target is observable in a given time step. I want to convert this to a list of time intervals.

Under the hood, both of these functions call a helper called _ezclump that has the following docstring:

    Return list of slices corresponding to the unmasked clumps of a 1-D array.
    (A "clump" is defined as a contiguous region of the array).

This is almost exactly what I want. (Really, I want to clump values along each row in a 2D array, and return a list of a list of slices... but this is close enough.)

This could be called np.flat_clump_nonzero or something similar.

Start by reading np.ma.clump_masked, np.ma.clump_unmasked, and their _ezclump helper, which the issue identifies as closely related entry points. Compare their 1-D behavior with the requested ordinary-array runs, then clarify the public API and 2-D row semantics. Done means the behavior and returned slice structure are specified and implemented.