Fix issue for pandera allowing generic Series to work #492
Conversation
| index: Axes | None = ..., | ||
| dtype=..., | ||
| name: Hashable | None = ..., | ||
| copy: bool = ..., | ||
| fastpath: bool = ..., | ||
| ) -> Series[IntervalT]: ... | ||
| ) -> Series[Interval[int]]: ... | ||
| @overload |
There was a problem hiding this comment.
might be nice to add a TODO referencing the mypy issue
There was a problem hiding this comment.
might be nice to add a TODO referencing the mypy issue
I didn't research whether there exists a mypy issue, or how to create a small sample to replicate.
I could add a comment indicating that we prefer the Series[IntervalT] notation, but that it doesn't seem to work and reference the original pandas-stubs issue. Let me know if I should do that.
There was a problem hiding this comment.
Would something like the following work?
T = TypeVar("T", float, int, Timestamp, Timedelta)
data: IntervalIndex[Interval[T]],
....
) -> Series[Interval[T]]: ...There was a problem hiding this comment.
I can't see how this would work, because we have:
class IntervalIndex(IntervalMixin, ExtensionIndex, Generic[IntervalT]):Since IntervalIndex is generic, and Interval is generic, you can't make IntervalIndex depend on the "inner" generic type.
|
Thanks @Dr-Irv ! |
* split Series[IntervalT] into separate __new__ * add test for generic series * Update reference to GH issue
test_series.py:test_pandera_generic()Fix here is to expand the
Series[IntervalT].__new__()into each individual part. Seems like a doubleTypeVarcauses issues formypy. Had to do a similar fix forInterval.__contains__()Either @twoertwein or @bashtage can approve or merge (both not required)