Item access in sptensor always returns zero about scikit-tensor

[mnick]

I can't reproduce the behaviour you are describing. Can you provide a test case for the bug? The following test seems to work correctly:

def test_getitem():
    subs = (
        array([0, 1, 0, 5, 7, 8]),
        array([2, 0, 4, 5, 3, 9]),
        array([0, 1, 2, 2, 1, 0])
    )
    vals = array([1, 2, 3, 4, 5, 6])
    S = sptensor(subs, vals, shape=[10, 10, 3])
    assert_equal(0, S[1, 1, 1])
    assert_equal(0, S[1, 2, 3])
    assert_equal(1, S[0, 2, 0])
    assert_equal(2, S[1, 0, 1])
    assert_equal(3, S[0, 4, 2])
    assert_equal(4, S[5, 5, 2])
    assert_equal(5, S[7, 3, 1])
    assert_equal(6, S[8, 9, 0])

[ucarbehlul]

Hmm your example is working indeed.

But I found that if I use python array instead of numpy array, it won't work. Since it works correctly with ndarray, the bug may be resolved but I think it may be helpful to add a notice for this to docstrings.

Try it like this:

def test_getitem():
    subs = (
        [0, 1, 0, 5, 7, 8],
        [2, 0, 4, 5, 3, 9],
        [0, 1, 2, 2, 1, 0]
    )
    vals = [1, 2, 3, 4, 5, 6]
    S = sptensor(subs, vals, shape=[10, 10, 3])
    assert_equal(0, S[1, 1, 1])
    assert_equal(0, S[1, 2, 3])
    assert_equal(1, S[0, 2, 0])
    assert_equal(2, S[1, 0, 1])
    assert_equal(3, S[0, 4, 2])
    assert_equal(4, S[5, 5, 2])
    assert_equal(5, S[7, 3, 1])
    assert_equal(6, S[8, 9, 0])

[kastnerkyle]

@mnick do you think adding support for lists as indices and values for sparse tensors is worthwhile? Basically something like

if type(subs) != np.ndarray:
    subs = np.array(subs)

and the same for vals. Or would it be better to raise an error if the type is not ndarray?

[mnick]

For now I would prefer the later until the API / functionality matured.