DOC: fix a small np.einsum example (numpy#25707)

* DOC: fix a small np.einsum example

"..." is required if `a` is multi-dimensional, e.g.:
a = np.arange(25).reshape(5,5)

Otherwise, this error is raised:
ValueError: operand has more dimensions than subscripts given in
einstein sum, but no '...' ellipsis provided to broadcast the extra
dimensions.

numpy/_core/_add_newdocs.py

@@ -2009,8 +2009,9 @@
     identifier '->' as well as the list of output subscript labels.
     This feature increases the flexibility of the function since
     summing can be disabled or forced when required. The call
-    ``np.einsum('i->', a)`` is like :py:func:`np.sum(a, axis=-1) <numpy.sum>`,
-    and ``np.einsum('ii->i', a)`` is like :py:func:`np.diag(a) <numpy.diag>`.
+    ``np.einsum('i->', a)`` is like :py:func:`np.sum(a) <numpy.sum>`
+    if ``a`` is a 1-D array, and ``np.einsum('ii->i', a)``
+    is like :py:func:`np.diag(a) <numpy.diag>` if ``a`` is a square 2-D array.
     The difference is that `einsum` does not allow broadcasting by default.
     Additionally ``np.einsum('ij,jh->ih', a, b)`` directly specifies the
     order of the output subscript labels and therefore returns matrix
@@ -2019,6 +2020,8 @@
     To enable and control broadcasting, use an ellipsis.  Default
     NumPy-style broadcasting is done by adding an ellipsis
     to the left of each term, like ``np.einsum('...ii->...i', a)``.
+    ``np.einsum('...i->...', a)`` is like 
+    :py:func:`np.sum(a, axis=-1) <numpy.sum>` for array ``a`` of any shape.
     To take the trace along the first and last axes,
     you can do ``np.einsum('i...i', a)``, or to do a matrix-matrix
     product with the left-most indices instead of rightmost, one can do

numpy/_core/einsumfunc.py

@@ -1180,17 +1180,19 @@ def einsum(*operands, out=None, optimize=False, **kwargs):
     identifier '->' as well as the list of output subscript labels.
     This feature increases the flexibility of the function since
     summing can be disabled or forced when required. The call
-    ``np.einsum('i->', a)`` is like
-    :py:func:`np.sum(a, axis=-1) <numpy.sum>`, and ``np.einsum('ii->i', a)``
-    is like :py:func:`np.diag(a) <numpy.diag>`. The difference is that
-    `einsum` does not allow broadcasting by default. Additionally
-    ``np.einsum('ij,jh->ih', a, b)`` directly specifies the order of
-    the output subscript labels and therefore returns matrix multiplication,
-    unlike the example above in implicit mode.
+    ``np.einsum('i->', a)`` is like :py:func:`np.sum(a) <numpy.sum>`
+    if ``a`` is a 1-D array, and ``np.einsum('ii->i', a)``
+    is like :py:func:`np.diag(a) <numpy.diag>` if ``a`` is a square 2-D array.
+    The difference is that `einsum` does not allow broadcasting by default. 
+    Additionally ``np.einsum('ij,jh->ih', a, b)`` directly specifies the
+    order of the output subscript labels and therefore returns matrix
+    multiplication, unlike the example above in implicit mode.
 
     To enable and control broadcasting, use an ellipsis.  Default
     NumPy-style broadcasting is done by adding an ellipsis
     to the left of each term, like ``np.einsum('...ii->...i', a)``.
+    ``np.einsum('...i->...', a)`` is like 
+    :py:func:`np.sum(a, axis=-1) <numpy.sum>` for array ``a`` of any shape.
     To take the trace along the first and last axes,
     you can do ``np.einsum('i...i', a)``, or to do a matrix-matrix
     product with the left-most indices instead of rightmost, one can do