feat: simplified type which maps to pandas timedelta / pyarrow duration semantics

[tswast]

Is your feature request related to a problem?

In BigQuery DataFrames, we'd like to provide an API to allow users to get something like a pandas timedelta / pyarrow duration for some datetime / timestamp operations. Currently in Ibis (and BigQuery for that matter) we have two options, neither of which is fully ideal:

1. Use BigQuery (and Ibis) "interval" data type. This is a superset of the timedelta/duration behavior in that it also allows for calendar-aware days & months, similar to the pyarrow month_day_nano_interval type. This type is a bit awkard to work with at the moment and doesn't really map to anything useful in pandas (yet?)
2. Use integers and keep track of the units on our own.

Describe the solution you'd like

I'd love for an ibis duration type similar to the pyarrow type. For backends that don't have an equivalent, pandas could keep track of the units locally.

What version of ibis are you running?

8.0.0

What backend(s) are you using, if any?

BigQuery (maybe polars in future?)

Code of Conduct

• I agree to follow this project's Code of Conduct

[tswast]

Related discussion: #8138 (comment) which details problems with the current interval type.