SIM401 triggers on dict like objects that aren't proper dictionaries

[jogo-openai]

Description

$ ruff --version
ruff 0.9.3

class Foo:
    def __init__(self):
        self._dict = {}

    def __getitem__(self, key):
        return self._dict[key]

    def __setitem__(self, key, value):
        self._dict[key] = value

    def __iter__(self):
        return self._dict.__iter__()



foo = Foo()
foo['key'] = 'value'


if "key" in foo:
    value = foo["key"]
else:
    value = None

print(value)


# print(foo.get('key', None))
# Triggers `AttributeError: 'Foo' object has no attribute 'get'`

ruff check --fix --unsafe-fixes --diff --select=SIM401 foo.py

--- foo.py
+++ foo.py
@@ -17,10 +17,7 @@
 foo['key'] = 'value'


-if "key" in foo:
-    value = foo["key"]
-else:
-    value = None
+value = foo.get('key', None)

 print(value)


Would fix 1 error.

[dylwil3]

I'm slightly in favor of just updating the documentation (which is part of #15584) for this unsafe fix. An alternative is to restrict to instances where our limited type inference can tell that we have an honest dict object, but that would be a significant restriction in applicability. That said, false positives are more annoying than false negatives for this sort of rule, so I could be convinced of the second route.

[jogo-openai]

I have a slight preference for removing false positives and making sure things are actually dicts. That being said, just updating the documentation for this unsafe fix also seems good to me (I suspect this means llarge code bases would opt to not use this rule).