TEST-#7643: Fix residual failures from pandas 2.3 (#7644)

This PR fixes 2 failures I missed while adding support for pandas 2.3:
- `DataFrame.insert` apparently is now typed with `pandas-stubs==2.3`,
but was not typed before. Since pandas 2.3 is compatible with Python 3.9
all CI runs should pick up this change, but if it continues to be an
issue I'll look into a workaround for seeing if it's possible to
conditionally add the `type: ignore` directive.
- `np.cumproduct` was
[deprecated](https://numpy.org/doc/1.26/release/1.25.0-notes.html#deprecations)
some time ago, and removed in numpy>2. We've supported numpy>2 for a
while, so it's unclear why this issue only just triggered.

Signed-off-by: Jonathan Shi <jonathan.shi@snowflake.com>

Author: sfc-gh-joshi

modin/core/dataframe/base/dataframe/utils.py

@@ -105,10 +105,12 @@ def join_columns(
             left_on = []
             right_on = []
         # in other cases, we can simply add the index name to columns and proceed as normal
+        # on python 3.9 with pandas-stubs 2.2, these lines will warn about insert being an untyped call,
+        # but this error is no longer present on higher versions
         elif left_on[0] not in left:
-            left = left.insert(loc=0, item=left_on[0])  # type: ignore
+            left = left.insert(loc=0, item=left_on[0])  # type: ignore[no-untyped-call, unused-ignore]
         elif right_on[0] not in right:
-            right = right.insert(loc=0, item=right_on[0])  # type: ignore
+            right = right.insert(loc=0, item=right_on[0])  # type: ignore[no-untyped-call, unused-ignore]
 
     if any(col not in left for col in left_on) or any(
         col not in right for col in right_on

modin/tests/pandas/dataframe/test_udf.py

@@ -130,7 +130,7 @@ def test_aggregate_error_checking():
         modin_df.aggregate({modin_df.columns[0]: "sum", modin_df.columns[1]: "mean"})
 
     with warns_that_defaulting_to_pandas():
-        modin_df.aggregate("cumproduct")
+        modin_df.aggregate("arcsin")
 
 
 @pytest.mark.parametrize(