pandas: handle setitem with Timedelta

[Dr-Irv]

pylance v2022.1

The following code:

import pandas as pd

df = pd.DataFrame(
    {
        "x": pd.to_datetime(["2022-01-01", "2022-01-02", "2022-01-03"]),
        "a": ["a", "b", "a"],
    }
)

df.loc[df.a == "a", "x"] += pd.Timedelta("1d")

produces

Argument of type "Timedelta" cannot be assigned to parameter "value" of type "Scalar | ndarray[Unknown, Unknown] | Series[Dtype@__setitem__] | DataFrame" in function "__setitem__"
  Type "Timedelta" cannot be assigned to type "Scalar | ndarray[Unknown, Unknown] | Series[Dtype@__setitem__] | DataFrame"
    "Timedelta" is incompatible with "str"
    "Timedelta" is incompatible with "bytes"
    "Timedelta" is incompatible with "date"
    "Timedelta" is incompatible with "datetime"
    "Timedelta" is incompatible with "timedelta"
    "Timedelta" is incompatible with "bool"
    "Timedelta" is incompatible with "int"

This PR fixes that issue.