(self)
| 1030 | assert_(dx.dtype == np.dtype('timedelta64[D]')) |
| 1031 | |
| 1032 | def test_masked(self): |
| 1033 | # Make sure that gradient supports subclasses like masked arrays |
| 1034 | x = np.ma.array([[1, 1], [3, 4]], |
| 1035 | mask=[[False, False], [False, False]]) |
| 1036 | out = gradient(x)[0] |
| 1037 | assert_equal(type(out), type(x)) |
| 1038 | # And make sure that the output and input don't have aliased mask |
| 1039 | # arrays |
| 1040 | assert_(x._mask is not out._mask) |
| 1041 | # Also check that edge_order=2 doesn't alter the original mask |
| 1042 | x2 = np.ma.arange(5) |
| 1043 | x2[2] = np.ma.masked |
| 1044 | np.gradient(x2, edge_order=2) |
| 1045 | assert_array_equal(x2.mask, [False, False, True, False, False]) |
| 1046 | |
| 1047 | def test_second_order_accurate(self): |
| 1048 | # Testing that the relative numerical error is less that 3% for |
nothing calls this directly
no test coverage detected