add type hints to base.py
#1302
Conversation
|
Is there any task remaining for this PR? |
- annotations import in all files using typehints - typehint specific imports under type_checking condition - native library imports > installed libs > module import
plexapi/base.py
Outdated
| from typing import (TYPE_CHECKING, Any, Callable, Dict, List, Optional, Set, | ||
| Type, TypeVar, Union, cast, overload) |
There was a problem hiding this comment.
Use individual lines (I want to eventually use Black formatting). Similar other files.
| from typing import (TYPE_CHECKING, Any, Callable, Dict, List, Optional, Set, | |
| Type, TypeVar, Union, cast, overload) | |
| from typing import ( | |
| TYPE_CHECKING, | |
| Any, | |
| Callable, | |
| Dict, | |
| List, | |
| Optional, | |
| Set, | |
| Type, | |
| TypeVar, | |
| Union, | |
| cast, | |
| overload, | |
| ) |
There was a problem hiding this comment.
default config or have something in mind for this project?
- all single quotes will be converted to double unless "--skip-string-normalization" is passed
- all docstrings whitespaces will be stripped
- should a pyproject.toml be created for black config?
plexapi/base.py
Outdated
| """ Builds the details key with the XML include parameters. | ||
| All parameters are included by default with the option to override each parameter | ||
| or disable each parameter individually by setting it to False or 0. | ||
| """ | ||
| details_key = self.key | ||
| details_key = str(self.key) |
There was a problem hiding this comment.
If self.key is None then str(None) turns into 'None' so you end up return 'None'.
There was a problem hiding this comment.
good catch, removed casting
side effect: d4fc723#r1452965661
| """ Returns True if this object is a child of the given attributes. | ||
| This will search the parent objects all the way to the top. | ||
|
|
||
| Parameters: | ||
| **kwargs (dict): The attributes and values to search for in the parent objects. | ||
| See all possible `**kwargs*` in :func:`~plexapi.base.PlexObject.fetchItem`. | ||
| """ | ||
| obj = self | ||
| obj = self # ? Why is this needed? |
There was a problem hiding this comment.
So you don't mutate self in the while loop.
There was a problem hiding this comment.
isn't object address copied here? or am i missing something?
class Car:
def __init__(self, name: str, color: str) -> None:
self.name = name
self.color = color
def _dont_mutate_me(self) -> None:
obj = self # only memory address of self is copied not the object
obj.name = "foo" # will mutate self.name
assert id(obj) == id(self) # True
def __str__(self) -> str:
return f"I am a {self.color} {self.name}."
def main() -> None:
car = Car(name="Tesla", color="red")
print(car) # I am a red Tesla.
car._dont_mutate_me()
print(car) # I am a red foo.
if __name__ == "__main__":
main()
|
I also created a new branch for |
- fix _INCLUDES attr by including an empty dict - add explicit return types for methods missing them
| if ekey is None: | ||
| raise BadRequest('ekey was not provided') |
There was a problem hiding this comment.
Why is guard clause removed?
There was a problem hiding this comment.
def fetchItems(
self,
ekey: Union[str, List[int]],
cls: Optional[Type[PlexObjectT]] = None,
container_start: Optional[int] = None,
container_size: Optional[int] = None,
maxresults: Optional[int] = None,
**kwargs: Any,
):ekey is never None as defined by the method, hence unnecessary check. could typehint to accept None but then calling fetchItems without passing ekey does not make sense for the method, None should be handled by the caller imo
| if not data: | ||
| return [] |
There was a problem hiding this comment.
.query can return None and .findItems does not accept None. so it should be handled in some way. returned empty list as .fetchItems returns a list
There was a problem hiding this comment.
on a second thought maybe returning result is a better choice instead of empty list? or should raise error? what is the preferred way to handle no date? other choice is to make query raise an error instead of returning None
| if data is None: | ||
| raise NotFound(f'Unable to find elem: {key=}') |
There was a problem hiding this comment.
as long as query could return None, handling of None should be done, it is done here as next step requires subscripting data
self._loadData(data[0]) # None is not subscriptable| self._server = server | ||
| self._data = data | ||
| self._initpath = initpath or self.key | ||
| self._parent = weakref.ref(parent) if parent is not None else None | ||
| self._details_key = None |
There was a problem hiding this comment.
being assigned after few lines as below and not used before that assignment and this removal, hence the statement has no effect
self._details_key = self._buildDetailsKey()| if not data: | ||
| raise NotFound(f'Unable to find elem: {key=}') |
There was a problem hiding this comment.
for elem in data: # None not iterable
|
should I break this PR further down? Maybe on a per function basis? That should make it easier for review as currently this change is maybe so huge that it is daunting to start reviewing, and multiple threads going on for each change. |
I'm considering whether it would be more beneficial to merge it into
@JonnyWong16, I'd appreciate your thoughts on this. Do you think the benefits of direct merging outweigh any potential risks? either way I will close this PR and start anew since this is too huge an undertaking for review. |
Description
Support mypy and/or pyright for the
base.pyrelated #1296
Type of change
Please delete options that are not relevant.
Checklist: