Test docstring examples

[bveeramani]

Description

Add a job named doctest to the test.yml workflow. The job should run doctest against the examples in the nums package.

Use case

#153 added docstrings to functions. Now that we have examples in our code, we should ensure that examples run correctly because:

1. Incorrect examples are frustrating for the user.
2. If we test examples, we'll be more confident about the correctness of NumS.

For example, the examples in #436 should be tested.

    Examples
    --------
    Convert a list into an array:
    >>> a = [1, 2]
    >>> nps.asarray(a).get()
    array([1, 2])

    Existing arrays are not copied:
    >>> a = nps.array([1, 2])
    >>> nps.asarray(a) is a
    True

    If `dtype` is set, array is copied only if dtype does not match:
    >>> a = nps.array([1, 2], dtype=np.float32)
    >>> nps.asarray(a, dtype=np.float32) is a
    True
    >>> nps.asarray(a, dtype=np.float64) is a
    False

Open Questions

• Which Python versions should we run doctests against?
• Which backends should we run doctests against?
• Should doctests be part of a new job, or should we add them to the existing pytest job? (In this case, we'd want to rename the pytest job)