Simplify key in dct and dct[key] to dct.get(key)

[janosh]

# bad
if "key" in dct and dct["key"]:
    ...

# good
if dct.get("key"):
    ...

[agnes-sharan]

Hi could I work on this?

[charliermarsh]

Hi @agnes-sharan -- thanks for asking! We need to make a decision as maintainers on whether to include this (and how it should be categorized) before opening it up for contributors. We'll label it as such and comment here when there's consensus.

[sbrugman]

Examples in the wild: https://grep.app/search?q=%20%22%28.%2B%29%22%20in%20%5Ba-z0-9%5D%2B%20and%20%5Ba-z0-9%5D%2B%5C%5B%22%28.%2B%29%22%5C%5D%3A&regexp=true

It would make sense to propose this rule in flake8-simplify. They already have a couple of dict-related rules, and then port it to ruff (if they are responsive).

[janosh]

The regex can be made more generic to increase the number of hits. The site doesn't seem to support backreferences (.+) in .+ and .+\[\1\] so might be some false positives in there.

[sbrugman]

Refurb might be a candidate to include this rule into: https://github.com/dosisod/refurb

In the cases that dct is a common dict, this syntax is clearly more readable (just one ref to dct and key)

Alternatively, this could be a RUF category rule.

[charliermarsh]

We can add this as a ruff rule if there's interest.

[harupy]

@charliermarsh can I work on this?

[charliermarsh]

Go for it!

[harupy]

How would you name this rule?

[charliermarsh]

We may need to limit this to boolean tests, since it returns None rather than False.

[charliermarsh]

I'm not sure, maybe like UnnecessaryDictIn? Easy to rename during review though.