Question

我已经为这个问题写了一些代码。（python27）

图表用冷冻集密钥和frozensets集合表示为字典：

sample_graph = {frozenset([7]): set([frozenset([4]), frozenset([5]), frozenset([3])]), frozenset([5]): set([frozenset([7]), frozenset([2]), frozenset([1])]), frozenset([3]): set([frozenset([7]), frozenset([4]), frozenset([2]), frozenset([1])]), frozenset([6]): set([frozenset([4]), frozenset([2]), frozenset([1])]), frozenset([4]): set([frozenset([6]), frozenset([7]), frozenset([3]), frozenset([1])]), frozenset([1]): set([frozenset([6]), frozenset([4]), frozenset([5]), frozenset([2]), frozenset([3])]), frozenset([2]): set([frozenset([6]), frozenset([5]), frozenset([3]), frozenset([1])])}

输出应该是只有两个节点的图形，这两个节点是图形中所有节点的冻结点。此时它会遇到KeyError。

def kargerMinCut(graph):
if len(graph) == 2:
    return graph
u = random.choice(graph.keys())   # u and v are frozensets, idea is that they form
v = random.choice(list(graph[u])) # a clique in a single frozenset
for node in graph:
    if node != u and node != v:
        links = graph[node]       
        if u in links or v in links:
            links.add(frozenset(tuple(u | v))) # combine u and v to form one link
            links.discard(u)                   # delete old links to u and v
            links.discard(v)            
            graph[node] = links
graph[u | v] = graph[u] | graph[v]             # new key for u and v 
del graph[u], graph[v]                         # u and v are no longer needed
return kargerMinCut(graph)

Answer 1

我认为问题可能是使用is关键字。请注意，在python中，is仅在两个参数引用 exact 相同对象（相当于C ++中的char* == char*）时才返回true。==运算符返回true如果内容相同（相当于C ++中的string == string）。

而不是is not尝试!=。

在python中遍历图形中的元素时，曾经遇到过这个相同的问题。：）

PS--另外，我将以下一行写为完整if：

links.add(frozenset(tuple(u | v))) if u in links or v in links else None

我在Python中的图形收缩算法表现奇怪

1 个答案: