Python Code Review Checklist — Python Guidelines 0.1.0 documentation

Code is blackened with black

flake8 has been run with no errors

mypy has been run with no errors

Function complexity problems have been resolved using the default
complexity index of flake8.

Important core code can be loaded in iPython, ipdb easily.

There is no dead code

Comprehensions or generator expressions are used in place of for
loops where appropriate

Comprehensions and generator expressions produce state but they do
not have side effects within the expression.

Use zip(), any(), all(), etc. instead of for loops where
appropriate

Functions that take as parameters and mutate mutable variables don’t
return these variables. They return None.

Return immutable copies of mutable types instead of mutating the
instances themselves when mutable types are passed as parameters with
the intention of returning a mutated version of that variable.

Avoid method cascading on objects with methods that return self.

Function and method parameters never use empty collection or sequence
instances like list [] or dict {}. Instead they must use
None to indicate missing input

Variables in a function body are initialised with empty sequences or
collections by callables, list(), dict(), instead of [],
{}, etc.

Always use the Final type hint for class instance parameters that
will not change.

Context-dependent variables are not unnecessarily passed between
functions or methods

View functions either implement the business rules the view is
repsonsible for or it passes data downstream to have this done by
services and receives non-context dependent data back.

View functions don’t pass request to called functions

Functions including class methods don’t have too many local
parameters or instance variables. Especially a class’ __init__()
should not have too many parameters.

Profiling code is minimal

Logging is the minimum required for production use

There are no home-brewed solutions for things that already exist in
the PSL