New signal typing #594
New signal typing #594
Conversation
|
Apologies, another unreviewable PR, but if you start from the tests you can see the changes needed |
| "numpy<2.0.0", | ||
| "numpy", |
There was a problem hiding this comment.
Do we not have this working with numpy >=2? That's what's being installed currently right?
There was a problem hiding this comment.
These changes will make use of np.dtypes.StringDType() as discussed with @danielballan which is only available in numpy 2, so yes I should make it numpy >= 2
| "packaging", | ||
| "pint", | ||
| "bluesky>=1.13.0a3", | ||
| "event_model", | ||
| "p4p", | ||
| "event-model @ git+https://github.com/bluesky/event-model@main", |
There was a problem hiding this comment.
Why do we want to pin main here? Using latest release would incentivize regular event-model patch releases.
There was a problem hiding this comment.
Ah I see you're using Limits
There was a problem hiding this comment.
There will probably be a few other import level changes to event model before I'm done, then we can release event model too and go back to a release here
| if not get_origin(datatype) == np.ndarray: | ||
| raise TypeError(f"Expected np.ndarray, got {datatype}") | ||
| # datatype = numpy.ndarray[typing.Any, numpy.dtype[numpy.float64]] | ||
| # so extract numpy.float64 from it | ||
| return np.dtype(get_args(get_args(datatype)[1])[0]) |
There was a problem hiding this comment.
Do we allow for np arrays without subtype? If so we need:
| if not get_origin(datatype) == np.ndarray: | |
| raise TypeError(f"Expected np.ndarray, got {datatype}") | |
| # datatype = numpy.ndarray[typing.Any, numpy.dtype[numpy.float64]] | |
| # so extract numpy.float64 from it | |
| return np.dtype(get_args(get_args(datatype)[1])[0]) | |
| if get_origin(datatype) == np.ndarray: | |
| # np.ndarray[typing.Any, np.dtype[np.float64]] returns np.float64 | |
| return np.dtype(get_args(get_args(datatype)[1])[0]) | |
| elif datatype == np.ndarray: | |
| return np.float64 | |
| else: | |
| raise TypeError(f"Expected np.ndarray, got {datatype}") |
If not then perhaps a modification to the error message that the numpy array must be explicitly typed in the npt.NDArray format?
Co-authored-by: Eva Lott <[email protected]>
Co-authored-by: Eva Lott <[email protected]>
Co-authored-by: Eva Lott <[email protected]>
WIP for review by @evalott100