TYP: offsets.pyi #45331
TYP: offsets.pyi #45331
Conversation
|
Hello @twoertwein! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found: There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻 Comment last updated at 2022-02-17 03:27:46 UTC |
| @@ -605,7 +605,8 @@ def asfreq(self, freq=None, how: str = "E") -> PeriodArray: | |||
|
|
|||
| freq = Period._maybe_convert_freq(freq) | |||
|
|
|||
| base1 = self.freq._period_dtype_code | |||
| # error: "BaseOffset" has no attribute "_period_dtype_code" | |||
| base1 = self.freq._period_dtype_code # type: ignore[attr-defined] | |||
There was a problem hiding this comment.
This error occurs a few times. BaseOffset does not have this attribute but almost all(?) sub-classes of BaseOffset have this attribute.
| # treat this specifically as timedelta-NaT | ||
| result = np.empty(self.shape, dtype=np.float64) | ||
| result.fill(np.nan) | ||
| return result | ||
|
|
||
| # dispatch to Timedelta implementation | ||
| result = other.__rfloordiv__(self._ndarray) | ||
| return result | ||
| return other.__rfloordiv__(self._ndarray) |
There was a problem hiding this comment.
Combined these two lines to not assign a new type to result
| @overload | ||
| def __add__(self, other: npt.NDArray[np.object_]) -> npt.NDArray[np.object_]: ... | ||
| @overload | ||
| def __add__(self: _BaseOffsetT, other: BaseOffset) -> _BaseOffsetT: ... |
There was a problem hiding this comment.
These operators are not "perfect" but they might be good enough for now. Based on the current annotations, __add__(self: Year, other: Second) is declared to return a Year but it actually returns Second
|
@twoertwein can you merge main |
|
@jbrockmendel rebased and green |
| @@ -219,7 +219,9 @@ def arithmetic_op(left: ArrayLike, right: Any, op): | |||
| # (https://github.com/pandas-dev/pandas/issues/41165) | |||
| _bool_arith_check(op, left, right) | |||
|
|
|||
| res_values = _na_arithmetic_op(left, right, op) | |||
| # error: Argument 1 to "_na_arithmetic_op" has incompatible type | |||
There was a problem hiding this comment.
likely avoidable if we inline should_extension_dispatch
| freq._period_dtype_code == own_freq._period_dtype_code | ||
| freq._period_dtype_code | ||
| # error: "BaseOffset" has no attribute "_period_dtype_code" | ||
| == own_freq._period_dtype_code # type: ignore[attr-defined] |
There was a problem hiding this comment.
i think this is a perf optimization for dtype == self.dtype that might no longer be necessary; i think we do this in PeriodDtype.eq now.
There was a problem hiding this comment.
Should probably be done in a separate PR.
|
@twoertwein status here? |
I think it should be ready. The open two comments would require non-typing changes, which I would like to avoid in this PR. |
|
thanks @twoertwein |
|
hmm maybe should have rebase first, oh well, let's see |
This PR and #44922 are the last parts of #43744.