Infer n_splits in split helper

pytensor/tensor/basic.py

@@ -2176,9 +2176,13 @@ def matrix_transpose(x: "TensorLike") -> TensorVariable:
     return swapaxes(x, -1, -2)
 
 
-def split(x, splits_size, n_splits, axis=0):
-    the_split = Split(n_splits)
-    return the_split(x, axis, splits_size)
+def split(x, splits_size, *, n_splits=None, axis=0):
+    if n_splits is None:
+        if isinstance(splits_size, Variable):
+            n_splits = get_vector_length(splits_size)
+        else:
+            n_splits = len(splits_size)
+    return Split(n_splits)(x, axis, splits_size)
 
 
 class Split(COp):

pytensor/tensor/fourier.py

@@ -116,8 +116,8 @@ class Fourier(Op):
             l = len(shape_a)
             shape_a = stack(shape_a)
             out_shape = concatenate((shape_a[0:axis], [n], shape_a[axis + 1 :]))
-            n_splits = [1] * l
-            out_shape = split(out_shape, n_splits, l)
+            splits = [1] * l
+            out_shape = split(out_shape, splits, n_splits=l)
             out_shape = [a[0] for a in out_shape]
         return [out_shape]
 

tests/link/mlx/test_core.py

@@ -148,7 +148,7 @@ def test_empty_dynamic_shape():
 def test_split_const_axis_const_splits_compiled():
     x = pt.vector("x")
     splits = [2, 3]
-    outs = pt.split(x, splits, len(splits), axis=0)
+    outs = pt.split(x, splits, n_splits=len(splits), axis=0)
     compare_mlx_and_py([x], outs, [np.arange(5, dtype="float32")])
 
 
@@ -156,7 +156,7 @@ def test_split_dynamic_axis_const_splits():
     x = pt.matrix("x")
     axis = pt.scalar("axis", dtype="int64")
     splits = [1, 2, 3]
-    outs = pt.split(x, splits, len(splits), axis=axis)
+    outs = pt.split(x, splits, n_splits=len(splits), axis=axis)
 
     test_input = np.arange(12).astype(config.floatX).reshape(2, 6)
 

tests/link/numba/test_tensor_basic.py

@@ -209,7 +209,7 @@ def test_Join(vals, axis):
 def test_Split(n_splits, axis, values, sizes):
     values, values_test = values
     sizes, sizes_test = sizes
-    g = pt.split(values, sizes, n_splits, axis=axis)
+    g = pt.split(values, sizes, n_splits=n_splits, axis=axis)
     assert len(g) == n_splits
     if n_splits == 0:
         return

tests/link/pytorch/test_basic.py

@@ -518,7 +518,7 @@ rng = np.random.default_rng(42849)
 def test_Split(n_splits, axis, values, sizes):
     i = pt.tensor("i", shape=values.shape, dtype=config.floatX)
     s = pt.vector("s", dtype="int64")
-    g = pt.split(i, s, n_splits, axis=axis)
+    g = pt.split(i, s, n_splits=n_splits, axis=axis)
     assert len(g) == n_splits
     if n_splits == 0:
         return