Thomas M Kehrenberg
Thomas M Kehrenberg
I just had a look at nptyping and it doesn't seem to define any type stubs for numpy functions (there are no .pyi in their repository). Where did you observe...
I think you need to first run ``` pip install -e . ``` in the main directory of the repository. I will add something about this to the README.
Do the errors look like this? ``` tests/pandas_test.py:3: error: Skipping analyzing 'pandas': found module but no type hints or library stubs tests/pandas_test.py:4: error: Skipping analyzing 'numpy': found module but no...
Yeah, that's a bug. I'm not sure we can fix it. Maybe we could use `List[Union[int, float]]`.
Hi! pandas indexing is indeed incredibly complex and a major source of headaches for this project. Currently, `.loc` indexing accepts these kinds of inputs: ```python _MaskType = Union[Series[bool], _np.ndarray[_np.bool_], List[bool]]...
Hmm, we observed something similar here: https://github.com/predictive-analytics-lab/data-science-types/pull/155#issuecomment-681001951 I'm not sure what causes this.
Without the setting, mypy doesn't check the type stubs at all.
> I'm assuming that without the virtualenv that the check_all does not throw these errors? If so, I wonder if this is a mypy issue. Yes, I don't get any...
Usually I use conda environments, so maybe it's related to that. I'll try a venv environment.