CI: Unpin MyPy #36012
CI: Unpin MyPy #36012
Conversation
simonjayhawkins
added
CI
| # NOTE: tokenize.Name is not a public constant | ||
| # error: Module has no attribute "Name" [attr-defined] | ||
| if not re.match("^" + tokenize.Name + "$", k): # type: ignore[attr-defined] | ||
| if not re.match("^" + tokenize.Name + "$", k): |
There was a problem hiding this comment.
latest typeshed (bundled with mypy) exposes undocumented attributes
There was a problem hiding this comment.
does this mean theres a lower bound for what mypy we want?
There was a problem hiding this comment.
I think that in general it is recommended to pin mypy see https://mypy.readthedocs.io/en/stable/existing_code.html?highlight=pin#continuous-integration
does this mean theres a lower bound for what mypy we want?
with warn_unused_ignores, yes.
There was a problem hiding this comment.
so I could change to mypy=0.782 (best practice) or mypy>=0.782 (see #27568 (comment))
There was a problem hiding this comment.
commented above, prefer either of these to no version
| # "Union[pandas.core.common.<subclass of "Iterable" and "Sized">, | ||
| # pandas.core.common.<subclass of "Iterable" and "Sized">1, T]", expected | ||
| # "Union[Collection[T], T]") [return-value] | ||
| obj = cast(Collection, obj) |
There was a problem hiding this comment.
our signature here is not entirely correct. If we pass an Iterable with a __len__ method which is not necessarily a collection, the object is returned unchanged.
I think best to leave signature unchanged for now and could use Protocol in the future.
There was a problem hiding this comment.
since i'm lying to the type checker a #type: ignore is probably more preferable to a cast.
| @@ -364,7 +364,7 @@ class BaseExprVisitor(ast.NodeVisitor): | |||
|
|
|||
| unary_ops = _unary_ops_syms | |||
| unary_op_nodes = "UAdd", "USub", "Invert", "Not" | |||
| unary_op_nodes_map = dict(zip(unary_ops, unary_op_nodes)) | |||
| unary_op_nodes_map = {k: v for k, v in zip(unary_ops, unary_op_nodes)} | |||
There was a problem hiding this comment.
mypy giving
error: Need type annotation for 'unary_op_nodes_map' (hint: "unary_op_nodes_map: Dict[, ] = ...") [var-annotated]
mypy is ok with a dictcomp, or could add Dict[str, str] as variable type annotation instead
| __setitem__ = __setslice__ = _disabled # type: ignore[assignment] | ||
| __delitem__ = __delslice__ = _disabled # type: ignore[assignment] | ||
| pop = append = extend = _disabled # type: ignore[assignment] | ||
| remove = sort = insert = _disabled # type: ignore[assignment] |
There was a problem hiding this comment.
we probably could make FrozenList Generic to allow type annotations such as FrozenList[Label]
This is an exercise for another day.
| [mypy-pandas.conftest] | ||
| ignore_errors=True | ||
|
|
||
| [mypy-pandas.tests.tools.test_to_datetime] |
There was a problem hiding this comment.
errors were fixed with latest typeshed
| ignore_errors=True | ||
|
|
||
| [mypy-pandas.tests.tools.test_to_datetime] | ||
| [mypy-pandas.conftest,pandas.tests.window.conftest] |
There was a problem hiding this comment.
similar error in pandas.tests.window.conftest as in pandas.conftest,pandas, so bundled together.
this is a mypy inference issue we have encountered before
error: List item 0 has incompatible type "ParameterSet"; expected "Sequence[Collection[object]]" [list-item]
|
LGTM |
|
LGTM |
environment.yml
Outdated
| @@ -21,7 +21,7 @@ dependencies: | |||
| - flake8-comprehensions>=3.1.0 # used by flake8, linting of unnecessary comprehensions | |||
| - flake8-rst>=0.6.0,<=0.7.0 # linting of code blocks in rst files | |||
| - isort>=5.2.1 # check that imports are in the right order | |||
| - mypy=0.730 | |||
| - mypy | |||
There was a problem hiding this comment.
would prefer we pin exactly for this, or at the very least a lower bound.
| # NOTE: tokenize.Name is not a public constant | ||
| # error: Module has no attribute "Name" [attr-defined] | ||
| if not re.match("^" + tokenize.Name + "$", k): # type: ignore[attr-defined] | ||
| if not re.match("^" + tokenize.Name + "$", k): |
There was a problem hiding this comment.
commented above, prefer either of these to no version
|
lgtm ex @jreback comments; definitely need to get this updated given all of the changes between 0.73 and now |
|
thanks @simonjayhawkins note may want to add this to the whatsnew for development changes (we usually list pytest version changes) |
No description provided.