Implementing repeat function

ci/Numba-array-api-xfails.txt

@@ -70,7 +70,7 @@ array_api_tests/test_creation_functions.py::test_empty_like
 array_api_tests/test_data_type_functions.py::test_finfo[complex64]
 array_api_tests/test_manipulation_functions.py::test_squeeze
 array_api_tests/test_has_names.py::test_has_names[utility-diff]
-array_api_tests/test_has_names.py::test_has_names[manipulation-repeat]
+
 array_api_tests/test_has_names.py::test_has_names[manipulation-tile]
 array_api_tests/test_has_names.py::test_has_names[manipulation-unstack]
 array_api_tests/test_has_names.py::test_has_names[statistical-cumulative_sum]
@@ -79,7 +79,7 @@ array_api_tests/test_has_names.py::test_has_names[indexing-take_along_axis]
 array_api_tests/test_has_names.py::test_has_names[searching-count_nonzero]
 array_api_tests/test_has_names.py::test_has_names[searching-searchsorted]
 array_api_tests/test_signatures.py::test_func_signature[diff]
-array_api_tests/test_signatures.py::test_func_signature[repeat]
+
 array_api_tests/test_signatures.py::test_func_signature[tile]
 array_api_tests/test_signatures.py::test_func_signature[unstack]
 array_api_tests/test_signatures.py::test_func_signature[take_along_axis]
@@ -107,7 +107,6 @@ array_api_tests/test_statistical_functions.py::test_cumulative_sum
 array_api_tests/test_array_object.py::test_getitem_arrays_and_ints_1[None]
 array_api_tests/test_array_object.py::test_getitem_arrays_and_ints_2[1]
 array_api_tests/test_array_object.py::test_getitem_arrays_and_ints_2[None]
-array_api_tests/test_manipulation_functions.py::test_repeat
 array_api_tests/test_searching_functions.py::test_count_nonzero
 array_api_tests/test_searching_functions.py::test_searchsorted
 array_api_tests/test_manipulation_functions.py::test_tile

sparse/numba_backend/__init__.py

@@ -119,6 +119,7 @@
     permute_dims,
     prod,
     real,
+    repeat,
     reshape,
     round,
     squeeze,
@@ -335,6 +336,7 @@
     "where",
     "zeros",
     "zeros_like",
+    "repeat",
 ]
 
 

sparse/numba_backend/_common.py

@@ -10,7 +10,7 @@
 
 import numpy as np
 
-from ._coo import as_coo
+from ._coo import as_coo, expand_dims
 from ._sparse_array import SparseArray
 from ._utils import (
     _zero_of_dtype,
@@ -3104,3 +3104,45 @@ def vecdot(x1, x2, /, *, axis=-1):
         x1 = np.conjugate(x1)
 
     return np.sum(x1 * x2, axis=axis, dtype=np.result_type(x1, x2))
+
+
+def repeat(a, repeats, axis=None):
+    """
+    Repeat each element of an array after themselves
+
+    Parameters
+    ----------
+    a : SparseArray
+        Input sparse arrays
+    repeats : int
+        The number of repetitions for each element.
+        (Uneven repeats are not yet Implemented.)
+    axis : int, optional
+        The axis along which to repeat values. Returns a flattened sparse array if not specified.
+
+    Returns
+    -------
+    out : SparseArray
+        A sparse array which has the same shape as a, except along the given axis.
+    """
+    if not isinstance(a, SparseArray):
+        raise TypeError("`a` must be a SparseArray.")
+
+    if not isinstance(repeats, int):
+        raise Exception("`repeats` must be an integer, uneven repeats are not yet Implemented.")
+
+    axes = list(range(a.ndim))
+    new_shape = list(a.shape)
+    axis_is_none = False
+    if axis is None:
+        a = a.reshape(-1)
+        axis = 0
+        axis_is_none = True
+    axes[a.ndim - 1], axes[axis] = axes[axis], axes[a.ndim - 1]
+    new_shape[axis] *= repeats
+    a = expand_dims(a, axis=axis + 1)
+    shape_to_broadcast = a.shape[: axis + 1] + (a.shape[axis + 1] * repeats,) + a.shape[axis + 2 :]
+    a = broadcast_to(a, shape_to_broadcast)
+    if not axis_is_none:
+        return a.reshape(new_shape)
+    return a.reshape(new_shape).flatten()

sparse/numba_backend/tests/test_coo.py

@@ -1926,3 +1926,29 @@ def test_xH_x():
     assert_eq(Ysp.conj().T @ Y, Y.conj().T @ Y)
     assert_eq(Ysp.conj().T @ Ysp, Y.conj().T @ Y)
     assert_eq(Y.conj().T @ Ysp.conj().T, Y.conj().T @ Y.conj().T)
+
+
+@pytest.mark.parametrize("ndim", range(1, 5))
+@pytest.mark.parametrize("repeats", [1, 2, 3])
+def test_repeat(ndim, repeats):
+    rng = np.random.default_rng()
+    shape = tuple(rng.integers(1, 4) for _ in range(ndim))
+    a = rng.integers(1, 10, size=shape)
+    with pytest.raises(TypeError, match="`a` must be a SparseArray"):
+        sparse.repeat(a, repeats=2)
+    sparse_a = COO.from_numpy(a)
+    with pytest.raises(Exception, match="`repeats` must be an integer"):
+        sparse.repeat(sparse_a, repeats=[2, 2, 2])
+    for axis in range(ndim):
+        expected = np.repeat(a, repeats=repeats, axis=axis)
+        result_sparse = sparse.repeat(sparse_a, repeats=repeats, axis=axis)
+        actual = result_sparse.todense()
+        assert actual.shape == expected.shape, f"Shape mismatch on axis {axis}: {actual.shape} vs {expected.shape}"
+        np.testing.assert_array_equal(actual, expected)
+
+    expected = np.repeat(a, repeats=repeats, axis=None)
+    result_sparse = sparse.repeat(sparse_a, repeats=repeats, axis=None)
+    actual = result_sparse.todense()
+    print(f"Expected: {expected}, Actual: {actual}")
+    assert actual.shape == expected.shape
+    np.testing.assert_array_equal(actual, expected)

sparse/numba_backend/tests/test_namespace.py

@@ -134,6 +134,7 @@ def test_namespace():
         "real",
         "reciprocal",
         "remainder",
+        "repeat",
         "reshape",
         "result_type",
         "roll",