"foldleft" support? about flupy HOT 2 CLOSED

The implementation of flu.reduce is based on functools.reduce.

from typeshed, python's functools.reduce supports both same and different types

def reduce(function: Callable[[T, T], T], sequence: Iterable[T]) -> T: 

def reduce(function: Callable[[S, T], S], sequence: Iterable[T], initial: S) -> S:

In contrast, flu.reduce only supports

I think the right solution would be to add

def fold_left(self: Fluent[T], function: Callable[[S, T], S], initial: S) -> S:

That is also implemented in terms of functools.reduce.

Based on SO reference below, the academic difference between the fold_left and reduce is that reduce assumes associativity. That assumption enables parallelism. Given that functools.reduce executes sequentially, they are interchangeable.

Assuming you agree with ^, Would you be interested in opening a PR? If not, I'm happy to knock it out within a few days.

References:

Explanation of reduce vs fold
https://stackoverflow.com/questions/25158780/difference-between-reduce-and-foldleft-fold-in-functional-programming-particula/25158790

Typeshed type definition of functools.reduce
https://github.com/python/typeshed/blob/d21370965aa2a1711d6986e509e43d651f2e3411/stdlib/3/functools.pyi#L10

Example implementation
https://www.burgaud.com/foldl-foldr-python

from flupy.

I've implemented it and opened a PR.
#12

from flupy.