Imagine a function which dynamically adds attributes to an object using setattr
. The reason for doing so is that I want to map some external structure (e.g. a given parameter tree) to an object:
my_object = SomeClass()
apply_structure(my_object, some_descriptor)
my_object.device1.enabled = True
Technically this works but of course Pylint rightly complains about 'device1' being not a member of SomeClass
.
I could disable the warning but that would be bad (because I still want to get the warning in all cases when the attribute does not exist because of misspelling, etc).
Is there a common and legal (Pylint-proof) way to dynamically add members to an object that not leads to warnings?
Alternatively: Can I disable Pylint for just one object rather than a line/block/file?
Explanation:
You might wonder why I should equip an object with member attributes dynamically when I plan to access these attributes in a hard-coded way later.
The reason is: I have a dynamic part of the program (where the decoration happens) and a static part which is specialized for a certain scenario. So I could also create a static class for this scenario but that would be overkill in a lot of situations.
The following specialized code might allow access to some parameter of a device which might be attached to some bus:
class MyDeviceHandler:
on_get_some_subtree_element(self):
return _some_internal_value
on_set_some_subtree_element(self, value):
_some_internal_value = value
dev = MyDeviceHandler()
decorate_object_with_device_structure(dev, 'some/attached/device')
dev.some.subtree.element = 5 <--- will call the set-callback
x = dev.some.subtree.element <--- will call the get-callback
So the structure behind 'some/attached/device'
might be arbitrary and very complex and I don't want to reproduce it in a class structure.
One way to get rid of this warning would be to create/access a dict
based tree:
dev['some']['subtree']['element'] = 5
But this is harder to write and not nice to read - I would only do this to quieten Pylint.
Just to provide the answer that works for me now - as The Compiler suggested you can add a rule for the problematic class in your projects .pylintrc
:
[TYPECHECK]
ignored-classes=Fysom,MyClass