Fix typing in context.py
#1471
Fix typing in context.py
#1471
Conversation
DanielNoord
added
the
Maintenance
astroid/context.py
Outdated
| self.args = args # Call positional arguments | ||
| if keywords: | ||
| keywords = [(arg.arg, arg.value) for arg in keywords] | ||
| keywords_list = [(arg.arg, arg.value) for arg in keywords] |
There was a problem hiding this comment.
I'm not a fan of putting the type in the variable name: it increase the mental load, you can check the type if you need it and the plural form indicate it's an iterable.
There was a problem hiding this comment.
I changed it to keywords_tuples. Open for any better suggestions.
There was a problem hiding this comment.
I changed it to
keywords_tuples. Open for any better suggestions.
I think that's worse, let's just leave it at keywords. It's assigned to self.keywords after all.
There was a problem hiding this comment.
I meant that keywords was fine :)
There was a problem hiding this comment.
Mypy will complain then. keywords is defined as keywords: Optional[List["Keyword"]] = None, on L161. That's why I introduced the name change.
There was a problem hiding this comment.
keywords: List[Tuple[str, str]] = [(arg.arg, arg.value) for arg in keywords]Will raise a no-redef warning..
There was a problem hiding this comment.
Hmm, maybe keywords_as_list then ? (or keywords_as_tuple)
There was a problem hiding this comment.
Hmm, maybe
keywords_as_listthen ?
They are a list already. Maybe keywords_list wasn't that bad after all.
There was a problem hiding this comment.
Went with arg_value_pairs 😅 Is that okay?
astroid/context.py
Outdated
| if TYPE_CHECKING: | ||
| from astroid.nodes.node_classes import Keyword, NodeNG | ||
|
|
||
| _InferenceCache = MutableMapping[ |
There was a problem hiding this comment.
Any idea why it was MutableMapping instead of Dict?
Do we want to start adding a Type suffix to type aliases?
There was a problem hiding this comment.
PR that added it is #1009.
Perhaps adding Type is something to do after we can run mypy successfully? I'm still running into issues with types not actually being correct or missing information. I think we can do a "cleanup" mypy PR after we can get it to pass.
There was a problem hiding this comment.
PR that added it is #1009.
Let's use Dict. Usually the return type should be precise whereas the parameter type should be open. Sequence["NodeNG"] is ok though I think. We assign a list converted to a tuple to it.
Perhaps adding
Typeis something to do after we can runmypysuccessfully? I'm still running into issues with types not actually being correct or missing information. I think we can do a "cleanup"mypyPR after we can get it to pass.
👍🏻
for more information, see https://pre-commit.ci
3 participants
Steps
Description
Had this laying around for quite some time. Mostly moving around some stuff and making sure
keywordsis not reassigned to another type.Type of Changes
Related Issue