timedelta_range or interval_range should accept Timedelta or timedelta as freq

[bemoody]

Describe the bug
The timedelta_range and interval_range functions should accept a pandas.Timedelta or datetime.timedelta object as the freq argument.

To Reproduce
Example:

import pandas

def foo(step: float):
    return pandas.timedelta_range(
        start=pandas.Timedelta(0),
        periods=10,
        freq=pandas.Timedelta(seconds=step),
    )

Using mypy foo.py:

foo.py:7: error: Argument "freq" to "timedelta_range" has incompatible type "Timedelta"; expected "Union[str, DateOffset, None]"  [arg-type]
Found 1 error in 1 file (checked 1 source file)

Please complete the following information:

• OS: Debian 11.6
• Python: 3.9.2
• mypy: 1.1.1
• pandas: 1.5.3
• pandas-stubs: 1.5.3.230321

Additional context
#223 mentions date_range and bdate_range.

pandas-dev/pandas#48631 updates the documentation for date_range, bdate_range, and interval_range.

#322 updates the stubs for date_range and bdate_range.

However, using Timedelta or timedelta with timedelta_range seems to work as well. (I'm not sure I fully understand the difference between Timedelta and DateOffset, but I don't see why a Timedelta shouldn't be accepted here.)

[Dr-Irv]

We generally follow a policy that the stubs follow what is documented for pandas.

So before we change the stubs, we should get agreement from the pandas maintainers that Timedelta and timedelta are supposed to be allowed with timedelta_range() and interval_range(). Can you open a doc issue in the pandas project there and cross-reference to here, so we can track that?

[bemoody]

The documentation has now been updated (thanks to @allisonkwan for the quick response.)

[Dr-Irv]

The documentation has now been updated (thanks to @allisonkwan for the quick response.)

Great. Now someone can do a PR to update the stubs here for date_range(), interval_range(), bdate_range(), and timedelta_range(), with appropriate tests.