Fix SIM222 and SIM223 false positives and auto-fix

[JonathanPlasse]

• Close SIM222 False positive #4058

Special case truthy expressions like side-effect expressions.
If a truthy expression is followed by the short-circuiting boolean, remove the short-circuiting boolean too as it is unnecessary.
The violation message should be changed in this case to something like "remove everything after the truth expression".

assert ("foo" and False) == "foo"
assert ("foo" or True) == "foo"

I use Truthiness from flake8-bandit, but flake8-pytest also has is_falsy_constant().
Should we make extract Truthiness from flake8-bandit, add the missing cases from flake8-pytest, and use it for flake8-bandit, flake8-pytest, and flake8-simplify?
I would be open making another PR for this.

[github-actions]

PR Check Results

Ecosystem

ℹ️ ecosystem check detected changes. (+0, -1, 0 error(s))

airflow (+0, -1)

- airflow/providers/apache/kafka/operators/consume.py:100:29: SIM222 [*] Use `True` instead of `... or True`

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.00     15.5±0.04ms     2.6 MB/sec    1.00     15.4±0.02ms     2.6 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.01      3.9±0.01ms     4.3 MB/sec    1.00      3.9±0.01ms     4.3 MB/sec
linter/all-rules/numpy/globals.py          1.03    431.4±8.32µs     6.8 MB/sec    1.00    420.7±1.34µs     7.0 MB/sec
linter/all-rules/pydantic/types.py         1.01      6.6±0.01ms     3.9 MB/sec    1.00      6.6±0.01ms     3.9 MB/sec
linter/default-rules/large/dataset.py      1.02      8.2±0.02ms     4.9 MB/sec    1.00      8.1±0.01ms     5.0 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.01   1791.0±3.17µs     9.3 MB/sec    1.00   1775.1±4.08µs     9.4 MB/sec
linter/default-rules/numpy/globals.py      1.00    184.6±0.93µs    16.0 MB/sec    1.00    185.5±1.35µs    15.9 MB/sec
linter/default-rules/pydantic/types.py     1.01      3.7±0.01ms     6.8 MB/sec    1.00      3.7±0.01ms     6.9 MB/sec
parser/large/dataset.py                    1.00      6.4±0.01ms     6.4 MB/sec    1.01      6.5±0.00ms     6.3 MB/sec
parser/numpy/ctypeslib.py                  1.00   1257.8±1.34µs    13.2 MB/sec    1.01   1272.6±1.50µs    13.1 MB/sec
parser/numpy/globals.py                    1.00    126.4±0.35µs    23.4 MB/sec    1.01    127.5±0.30µs    23.1 MB/sec
parser/pydantic/types.py                   1.00      2.8±0.00ms     9.3 MB/sec    1.01      2.8±0.00ms     9.2 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.01     17.6±0.33ms     2.3 MB/sec    1.00     17.4±0.29ms     2.3 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      4.4±0.07ms     3.8 MB/sec    1.02      4.5±0.08ms     3.7 MB/sec
linter/all-rules/numpy/globals.py          1.00   546.1±11.96µs     5.4 MB/sec    1.00   548.7±10.70µs     5.4 MB/sec
linter/all-rules/pydantic/types.py         1.00      7.3±0.10ms     3.5 MB/sec    1.01      7.4±0.11ms     3.4 MB/sec
linter/default-rules/large/dataset.py      1.00      8.8±0.06ms     4.6 MB/sec    1.03      9.1±0.14ms     4.5 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1925.1±18.13µs     8.6 MB/sec    1.03  1976.8±26.42µs     8.4 MB/sec
linter/default-rules/numpy/globals.py      1.00    216.5±3.16µs    13.6 MB/sec    1.03   223.7±13.01µs    13.2 MB/sec
linter/default-rules/pydantic/types.py     1.00      4.0±0.06ms     6.4 MB/sec    1.02      4.1±0.06ms     6.2 MB/sec
parser/large/dataset.py                    1.00      7.0±0.07ms     5.8 MB/sec    1.01      7.0±0.04ms     5.8 MB/sec
parser/numpy/ctypeslib.py                  1.00  1326.4±21.77µs    12.6 MB/sec    1.01  1335.3±15.68µs    12.5 MB/sec
parser/numpy/globals.py                    1.00    131.9±2.01µs    22.4 MB/sec    1.01    133.7±2.12µs    22.1 MB/sec
parser/pydantic/types.py                   1.00      3.0±0.04ms     8.4 MB/sec    1.00      3.0±0.03ms     8.4 MB/sec

[charliermarsh]

Yeah, I'd vote to extract that concept into ruff_python_semantic and use it in more places!

[JonathanPlasse]

I added in_test to Context.
It is considered to be in a test context when you are inside an if or an assert test, or inside the builtin bool.
Boolean operations inside other expressions than boolean operations are not in a test context.

For example:

# Inside test `a` is simplified

bool(a or [1] or True or [2])

assert a or [1] or True or [2]

if (a or [1] or True or [2]) and (a or [1] or True or [2]):
    pass

# Outside test `a` is not simplified

a or [1] or True or [2]

if (a or [1] or True or [2]) == (a or [1]):
    pass

if f(a or [1] or True or [2]):
    pass

[JonathanPlasse]

There is a false positive that has been fixed in airflow.

self.max_messages = max_messages or True

[JonathanPlasse]

Should I split this PR?

[charliermarsh]

I think it would be helpful. What do you think we can carve out into separate PRs?

[JonathanPlasse]

Here is the list of possible PRs

1. Move Truthiness into ruff_python_ast #4071
2. Add in_boolean_test to Context #4072
3. Fix SIM222 and SIM223 false positives and auto-fix #4063 (This PR)

I will rebase this PR on main after the first two are merged.

[JonathanPlasse]

The rebase is done.
It is ready for review.

[JonathanPlasse]

I will document the rules as we diverge a lot from the behavior of flake8-simplify.

[JonathanPlasse]

I documented the rules and detect when ... or True, True or ..., or ... or True or ... is used and adapted the violation message.

[JonathanPlasse]

You for your edits, it is much cleaner and clearer now.