Graphviz - 对齐节点和集群

Question

我正在尝试用 Graphviz 绘制一个有限状态图，但我无法像我想要的那样得到它。这是我到目前为止所写的内容:

digraph G {
    rankdir = LR;

    subgraph cluster_op1 {

        subgraph cluster_0 {
            1 -> 2 [label="a"]  
        }

        subgraph cluster_1 {
            3 -> {4 6} [label="ε"]
            4 -> 5 [label="b"]
            5 -> 4 [label="ε"]
            5 -> 6 [label="ε"]
        }

        subgraph cluster_2 {
            7 -> {8 10} [label="ε"]
            8 -> 9 [label="c"]
            9 -> 8 [label="ε"]
            9 -> 10 [label="ε"]
        }   

        2 -> 3 [label="ε"]
        6 -> 7 [label="ε"]
    }

    subgraph cluster_op2 {

        subgraph cluster_3 {
            11 -> {12 14} [label="ε"]
            12 -> 13 [label="ab"]
            13 -> 12 [label="ε"]
            13 -> 14 [label="ε"]
        }   

        subgraph cluster_4 {
            15 -> 16 [label="c"]    
        }

        14->15 [label="ε"]
    }

    0 -> {1 11} [label="ε"]
    {10 16} -> 17 [label="ε"]
}

这是它的样子:


我想要的是:

cluster_op1内的所有节点水平对齐(与 cluster_op2 相同)

cluster_op2居中于 cluster_op1

来自节点 0 的边至 cluster_op1和 cluster_op2长度相同(从这两个簇到节点 17 的边相同)

Answer 1

这不仅仅是一个真正的解决方案，但它为您提供了您想要的东西，至少对于这里介绍的案例。

它使用三个元素:

群 为了促进节点之间的直线，

重量 如果组因一个节点指向其他两个节点而混淆，所有这些节点都在同一组中；并使集群在它们之间对齐，

invis 边缘对齐 cluster_op2下面 cluster_op1 - 所以它不是居中而是手动调整。

另外，我已经更换了b -> a与 a -> b[ dir = back ]在适用的情况下；我认为这是避免难以发现的错误的好习惯。

所以这是您编辑的代码:

digraph G {
    rankdir = LR;
    0, 1, 2, 15, 16, 17
    3, 4, 5, 6     [ group = 1 ];
    7, 8, 9, 10    [ group = 2 ];
    11, 12, 13, 14 [ group = 3 ];
    2 -> 11[ style = invis ];

    subgraph cluster_op1 {

        subgraph cluster_0 {
            1 -> 2 [ label="a" ]; 
        }

        subgraph cluster_1 {
            3 -> 4 [ label="ε", weight = 4 ];
            3 -> 6 [ label="ε" ];
            4 -> 5 [ label="b" ];
            4 -> 5 [ label="ε", dir = back];
            5 -> 6 [ label="ε", weight = 4 ];
        }

        subgraph cluster_2 {
            7 -> {8 10} [label="ε"]
            8 -> 9 [label="c"]
            8 -> 9 [ label="ε", dir = back ];
            9 -> 10 [label="ε"]
        }   

        2 -> 3 [ label="ε", weight = 10 ;]
        6 -> 7 [ label="ε" ];
    }

    subgraph cluster_op2 {

        subgraph cluster_3 {
            11 -> 12 [ label="ε", weight = 4 ];
            11 -> 14 [ label="ε" ];
            12 -> 13 [ label="ab" ];
            12 -> 13 [ label="ε", dir = back ];
            13 -> 14 [ label="ε", weight = 4 ];
        }   

        subgraph cluster_4 {
            15 -> 16 [label="c"]    
        }

        14->15 [ label="ε", weight = 10 ];
    }

    0 -> {1 11} [label="ε"]
    {10 16} -> 17 [label="ε"]
}

这产生