wemake-python-styleguide: Forbid slow `in` inside comprehensions

Rule request

Thesis

len with comprehensions

Iterators have no len, and sometimes I forgetting it.

Bad:

len(1 for el in a if el in b)

Better:

len([1 for el in a if el in b])

`in` inside comprehensions

Good:

sum(1 for el in a if el in b)

Twice slower, but also ok:

sum(el in b for el in a)

Reasoning

Detect runtime TypeError in advance. We could also detect len from yield-like iterators, but resolving symbols in python always is a non-trivial thing, unfortunately.

About this issue

Original URL
State: open
Created 5 years ago
Comments: 16 (14 by maintainers)

Most upvoted comments

Would like to take this! Can u plz assign this to me

ManishAradwad on Oct 9, 2019

@ManishAradwad you need to create a new violation in best_practices, then we can create a new visitor here: https://github.com/wemake-services/wemake-python-styleguide/blob/master/wemake_python_styleguide/visitors/ast/loops.py

We need to visit:

ast.ListComp
ast.SetComp
ast.DictComp
ast.GeneratorExp

You can add a new method in the visitor: self._check_slow_in_expression(node)

Here’s how our bad node ((a > 0 for a in elements)) looks like (just one example):

GeneratorExp(elt=Compare(left=Name(id='a', ctx=Load(), lineno=1, col_offset=4), ops=[Gt()], comparators=[Num(n=0, lineno=1, col_offset=8)], lineno=1, col_offset=4), generators=[comprehension(target=Name(id='a', ctx=Store(), lineno=1, col_offset=14), iter=Name(id='elements', ctx=Load(), lineno=1, col_offset=19), ifs=[], is_async=0)], lineno=1, col_offset=4)

Then you write the required logic, test it, and submit a PR. I am here to help.

sobolevn on Oct 19, 2019