sweetmatch

This is sugar coating around the function scipy.optimize.linear_sum_assignment (Hungarian algo). The single provided function is sweetmatch.match, which takes 2 lists of any objects, a callable that evaluates how bad 2 objects match together (cost function) and an optional threshold value. The match function returns a list of matched pairs. Typing is correctly handled. In action:

from sweetmatch import match

x = [1, 2, 3]
y = [1.6, 2.1, 3.2]

def cost_function(x:int, y:float) -> float:
    return abs(x - y)

matches = match(x, y, cost_function)

print(matches)

# [(1, 1.6), (2, 2.1), (3, 3.2)]


matches = match(x, y, cost_function, cost_threshold=0.5)

print(matches)

# [(2, 2.1), (3, 3.2)]

In this example, the typing of the outputed matches is correctly handled and is list[tuple[int, float]].

Installation

pip install sweetmatch