accept float values in pd.Timedelta (#1026)

* allow float values in pd.Timedelta

closes #1015

* int is subtype of float

Author: 5j9

pandas-stubs/_libs/tslibs/timedeltas.pyi

@@ -102,7 +102,7 @@ class Timedelta(timedelta):
     value: int
     def __new__(
         cls,
-        value: str | int | Timedelta | timedelta | np.timedelta64 = ...,
+        value: str | float | Timedelta | timedelta | np.timedelta64 = ...,
         unit: TimeDeltaUnitChoices = ...,
         *,
         days: float | np.integer | np.floating = ...,

tests/test_timefuncs.py

@@ -149,6 +149,11 @@ def test_timedelta_series_arithmetic() -> None:
     r4: pd.TimedeltaIndex = tds1 / 10.2
 
 
+def test_timedelta_float_value() -> None:
+    # GH 1015
+    check(assert_type(pd.Timedelta(1.5, "h"), pd.Timedelta), pd.Timedelta)
+
+
 def test_timedelta_series_string() -> None:
     seq_list = ["1 day"]
     check(assert_type(pd.to_timedelta(seq_list), pd.TimedeltaIndex), pd.TimedeltaIndex)