Modify atleast_Nd to accept only one positional argument

pytensor/tensor/basic.py

@@ -4355,28 +4355,22 @@ def empty_like(
 
 
 def atleast_Nd(
-    *arys: np.ndarray | TensorVariable, n: int = 1, left: bool = True
+    arry: np.ndarray | TensorVariable, *, n: int = 1, left: bool = True
 ) -> TensorVariable:
-    """Convert inputs to arrays with at least `n` dimensions."""
-    res = []
-    for ary in arys:
-        ary = as_tensor(ary)
+    """Convert input to an array with at least `n` dimensions."""
 
-        if ary.ndim >= n:
-            result = ary
-        else:
-            result = (
-                shape_padleft(ary, n - ary.ndim)
-                if left
-                else shape_padright(ary, n - ary.ndim)
-            )
+    arry = as_tensor(arry)
 
-        res.append(result)
-
-    if len(res) == 1:
-        return res[0]
+    if arry.ndim >= n:
+        result = arry
     else:
-        return res
+        result = (
+            shape_padleft(arry, n - arry.ndim)
+            if left
+            else shape_padright(arry, n - arry.ndim)
+        )
+
+    return result
 
 
 atleast_1d = partial(atleast_Nd, n=1)

tests/tensor/test_basic.py

@@ -4364,7 +4364,8 @@ def test_atleast_Nd():
 
     for n in range(1, 3):
         ary1, ary2 = dscalar(), dvector()
-        res_ary1, res_ary2 = atleast_Nd(ary1, ary2, n=n)
+        res_ary1 = atleast_Nd(ary1, n=n)
+        res_ary2 = atleast_Nd(ary2, n=n)
 
         assert res_ary1.ndim == n
         if n == ary2.ndim: