sparse doesn't support shape kwarg in full_like

[max-sixty]

Describe the bug
I'm attempting to use NEP18 functionality to include sparse arrays in xarray's unstack refactor: pydata/xarray#4746

I'm not that familiar with the NEP18 interface, but currently I'm getting failures when using np.full_like and including a sparse keyword argument. Even if sparse doesn't require it, could it pass it through to maintain compatibility?

To Reproduce

In [1]: import sparse

In [2]: sparse.ones(5)
Out[2]: <COO: shape=(5,), dtype=float64, nnz=0, fill_value=1.0>

In [3]: x = _

In [4]: import numpy as np

In [12]: np.full_like(x, fill_value=0, shape=(2,2))
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-12-cf45b2f0c703> in <module>
----> 1 np.full_like(x, fill_value=0, shape=(2,2))

<__array_function__ internals> in full_like(*args, **kwargs)

/usr/local/lib/python3.8/site-packages/sparse/_sparse_array.py in __array_function__(self, func, types, args, kwargs)
    244             pass
    245         else:
--> 246             return sparse_func(*args, **kwargs)
    247
    248         try:

Expected behavior
A clear and concise description of what you expected to happen.

System

• OS and version: MacOS
• sparse version (sparse.__version__) 0.11.2
• NumPy version (np.__version__) 1.19.2
• Numba version (numba.__version__) 0.51.2

Additional context
Add any other context about the problem here.

[hameerabbasi]

This is fixed in the latest master.

[max-sixty]

Thanks @hameerabbasi