0

我在这里修改算法：

https://eli.thegreenplace.net/2018/unification/

我只有一组字符串和数字的等式。例如对于以下集合 l = [('x1', 'x2'), ('x2', 'x3')] 我希望映射说例如 {x1:x2, x2:x2, x3:x2 }。相反，我得到{'x1'：'x2'，'x2'：'x3'}。

这是修改后的代码。我不担心发生检查或除常量和字符串（变量）之外的任何其他结构。

def unify(x, y, subst):
    """Unifies term x and y with initial subst.

    Returns a subst (map of name->term) that unifies x and y, or None if
    they can't be unified. Pass subst={} if no subst are initially
    known. Note that {} means valid (but empty) subst.
    """
    if subst is None:
        return None
    elif x == y:
        return subst
    elif isinstance(x, str):
        return unify_variable(x, y, subst)
    elif isinstance(y, str):
        return unify_variable(y, x, subst)
    else:
        return None


def unify_variable(v, x, subst):
    """Unifies variable v with term x, using subst.
    Returns updated subst or None on failure.
    """
    assert isinstance(v, str)
    if v in subst:
        return unify(subst[v], x, subst)
    elif isinstance(x, str) and x in subst:
        return unify(v, subst[x], subst)
    else:
        return {**subst, v: x}



l = [('x1', 'x2'), ('x2', 'x3')]

subst = {}
for c in l:
    subst = unify(c[0], c[1], subst)

结果：{'x1': 'x2', 'x2': 'x3'}

有没有办法修改它以引用此映射范围内的相同变量？

4

0 回答 0