|
| 1 | +import random |
| 2 | + |
| 3 | +import numpy as np |
| 4 | +import pytest |
| 5 | + |
| 6 | +import arrayfire_wrapper.dtypes as dtypes |
| 7 | +from arrayfire_wrapper.lib._constants import Pad |
| 8 | +from arrayfire_wrapper.lib.create_and_modify_array.create_array.constant import constant |
| 9 | +from arrayfire_wrapper.lib.create_and_modify_array.create_array.pad import pad |
| 10 | +from arrayfire_wrapper.lib.create_and_modify_array.manage_array import get_dims |
| 11 | + |
| 12 | + |
| 13 | +@pytest.mark.parametrize( |
| 14 | + "original_shape", |
| 15 | + [ |
| 16 | + (random.randint(1, 100),), |
| 17 | + (random.randint(1, 100), random.randint(1, 100)), |
| 18 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 19 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 20 | + ], |
| 21 | +) |
| 22 | +def test_zero_padding(original_shape: tuple) -> None: |
| 23 | + """Test if pad creates an array with no padding if no padding is given""" |
| 24 | + original_array = constant(2, original_shape, dtypes.s64) |
| 25 | + padding = Pad(0) |
| 26 | + |
| 27 | + zero_shape = tuple(0 for _ in range(len(original_shape))) |
| 28 | + result = pad(original_array, zero_shape, zero_shape, padding) |
| 29 | + |
| 30 | + assert get_dims(result)[0:len(original_shape)] == original_shape |
| 31 | + |
| 32 | + |
| 33 | +@pytest.mark.parametrize( |
| 34 | + "original_shape", |
| 35 | + [ |
| 36 | + (random.randint(1, 100),), |
| 37 | + (random.randint(1, 100), random.randint(1, 100)), |
| 38 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 39 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 40 | + ], |
| 41 | +) |
| 42 | +def test_negative_padding(original_shape: tuple) -> None: |
| 43 | + """Test if pad can properly handle if negative padding is given""" |
| 44 | + with pytest.raises(RuntimeError): |
| 45 | + original_array = constant(2, original_shape, dtypes.s64) |
| 46 | + padding = Pad(0) |
| 47 | + |
| 48 | + neg_shape = tuple(-1 for _ in range(len(original_shape))) |
| 49 | + result = pad(original_array, neg_shape, neg_shape, padding) |
| 50 | + |
| 51 | + assert get_dims(result)[0:len(original_shape)] == original_shape |
| 52 | + |
| 53 | + |
| 54 | +@pytest.mark.parametrize( |
| 55 | + "original_shape", |
| 56 | + [ |
| 57 | + (random.randint(1, 100),), |
| 58 | + (random.randint(1, 100), random.randint(1, 100)), |
| 59 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 60 | + (random.randint(1, 100), random.randint(1, 100), random.randint(1, 100), random.randint(1, 100)), |
| 61 | + ], |
| 62 | +) |
| 63 | +def test_padding_shape(original_shape: tuple) -> None: |
| 64 | + """Test if pad outputs the correct shape when a padding is adding to the original array""" |
| 65 | + original_array = constant(2, original_shape, dtypes.s64) |
| 66 | + padding = Pad(0) |
| 67 | + |
| 68 | + beg_shape = tuple(random.randint(1, 10) for _ in range(len(original_shape))) |
| 69 | + end_shape = tuple(random.randint(1, 10) for _ in range(len(original_shape))) |
| 70 | + |
| 71 | + result = pad(original_array, beg_shape, end_shape, padding) |
| 72 | + new_shape = np.array(beg_shape) + np.array(end_shape) + np.array(original_shape) |
| 73 | + |
| 74 | + assert get_dims(result)[0:len(original_shape)] == tuple(new_shape) |
0 commit comments