No overload variant of "__sub__" of "Series" matches argument type "Timedelta"

[WillAyd]

Describe the bug
After upgrading to pandas 2.0 started getting errors as described in the title

To Reproduce

import pandas as pd

df = pd.read_csv("afile.csv", parse_dates=["foo"])
df["foo"] = df["foo"] - pd.Timedelta("1D")

Using mypy:

pandas-stubs-mre.py:4: error: No overload variant of "__sub__" of "Series" matches argument type "Timedelta"  [operator]
pandas-stubs-mre.py:4: note: Possible overload variants:
pandas-stubs-mre.py:4: note:     def __sub__(self, Timestamp | datetime | TimestampSeries, /) -> TimedeltaSeries
pandas-stubs-mre.py:4: note:     def __sub__(self, complex | ExtensionArray | ndarray[Any, Any] | dict[str, ndarray[Any, Any]] | list[Any] | tuple[Any, ...] | Index | Series[Any], /) -> Series[Any]
Found 1 error in 1 file (checked 1 source file)

Please complete the following information:

• OS: Discovered on both Windows / Linux
• python version: 3.11.0
• version of type checker: mypy 1.4.1 (compiled: yes)
• version of installed pandas-stubs: 2.0.2.230605

Additional context
The issue does not happen with older versions of pandas-stubs - 1.5.3.230321 issues no complaint. I haven't done a full bisect but guess this may go back to #616 ?

[Dr-Irv]

Thanks @WillAyd . This was bugging me too. I think there was an intermediate version of mypy that required #616 but now I put that code back in #759 and my local tests passed.