DOK indexing bug

[sayandip18]

Describe the bug
Indexing DOK with slices does not give the expected results

To Reproduce

>>> import sparse; shape=(3,3); density=0.11111111111111119; indices=(slice(1, None, -1), slice(None, None, 1))
>>> s = sparse.random(shape, density, format="dok")
>>> x = s.todense()
>>> x
array([[0.        , 0.        , 0.        ],
       [0.        , 0.        , 0.32402132],
       [0.        , 0.        , 0.        ]])
>>> s
<DOK: shape=(3, 3), dtype=float64, nnz=1, fill_value=0.0>
>>> s.todense()
array([[0.        , 0.        , 0.        ],
       [0.        , 0.        , 0.32402132],
       [0.        , 0.        , 0.        ]]) 
>>> s[indices]
<DOK: shape=(2, 3), dtype=float64, nnz=0, fill_value=0.0>
>>> s[indices].todense()
array([[0., 0., 0.],
       [0., 0., 0.]])
>>> x[indices]
array([[0.        , 0.        , 0.32402132],
       [0.        , 0.        , 0.        ]])

Expected behavior
s[indices][0,2] should be 0.32402132

System

• OS and version: Ubuntu 20.04.1
• sparse version: master