functools

Module nustd.functools.

Nu surface for Python's functools.

Only reduce is modeled: it is the one functools member that is a runtime value operation (a fold over a stream). The rest are out of Nu's value model and intentionally absent:

• partial / partialmethod / cmp_to_key produce callables - Nu has no first-class function value at the user surface.
• lru_cache / cache / cached_property are stateful (memoization) - they need the effect model (not yet built).
• wraps / update_wrapper / total_ordering are decorators over Python metadata, not runtime operations.

Import like the stdlib:

from nustd.functools import reduce

Call

reduce

Fold iterable left-to-right with function (functools.reduce).

functools.reduce(function, iterable, initializer=<UNSET>)

Path nustd.functools.reduce. Defined on nustd.functools.functions, bound as a function. Builds Object.

function is a lambda over the accumulator and the current item, called once at construction with refs to them, so it builds the step and never computes with them in Python. Given as a tree instead, it reads them via nu.Attr("acc") and nu.Attr("item"). With initializer the accumulator starts there; otherwise at the first item.

Arguments

Examples

nu.run(nustd.functools.reduce(lambda acc, x: nu.Int(acc) * 10 + x, [1, 2, 3]))[0]

123

The same as a tree:

nu.run(nustd.functools.reduce(nu.Int(nu.Attr("acc")) * 10 + nu.Attr("item"), [1, 2, 3]))[0]

123