ENH: create a way to take random uniform datetimes

Describe the issue:

Internally, the datetime64 dtype is represented as an integer. One can do operations on it such as adding or substracting, which are well defined.

Nevertheless, np.random.Generator.uniform will not let me use this dtype to specify the highs and lows. I would want to do something like this:

import numpy as np
low = np.datetime64("2020-01-01 00:00:00")
high = np.datetime64("2020-01-31 23:59:59")
np.random.default_rng().uniform(
    low,
    high,
    size=10
)

But it errors out so I have to add boilerplate code:

low + np.random.default_rng().uniform(
    high=(high - low).astype(int),
    size=10
).astype(int)

One could argue that in theory uniform is not the most appropriate choice since it generates floating points which then have to be rounded to integers, but np.random.Generator.integers also errors out:

np.random.default_rng().integers(low, high, size=10)

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-32-8d188f6ec7e3> in <module>
----> 1 np.random.default_rng().integers(low, high, size=10)

_generator.pyx in numpy.random._generator.Generator.integers()

_bounded_integers.pyx in numpy.random._bounded_integers._rand_int64()

TypeError: int() argument must be a string, a bytes-like object or a number, not 'datetime.datetime'

Reproduce the code example:

import numpy as np
low = np.datetime64("2020-01-01 00:00:00")
high = np.datetime64("2020-01-31 23:59:59")
np.random.default_rng().uniform(
    low,
    high,
    size=10
)

Error message:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-19-3abdd025d9c8> in <module>
      2 low = np.datetime64("2020-01-01 00:00:00")
      3 high = np.datetime64("2020-01-31 23:59:59")
----> 4 np.random.default_rng().uniform(
      5     low,
      6     high,

_generator.pyx in numpy.random._generator.Generator.uniform()

TypeError: float() argument must be a string or a number, not 'datetime.datetime'

NumPy/Python version information:

1.21.5 3.9.12 (main, Apr 5 2022, 06:56:58)
[GCC 7.5.0]

Start with the np.random.Generator.uniform and Generator.integers entry points shown in the traceback, then inspect the surrounding random-generation implementation and tests. Define how the requested datetime64 bounds should be handled and add coverage for the examples; the work is done when datetime64 inputs produce random datetime results without the conversion boilerplate.