I have the following the dot file contents:
digraph G {
start -> {a0, b0} -> end;
start -> c0 -> c1 -> c2 -> end;
start -> d0 -> d1 -> d2 -> end;
start -> {e0, f0} -> end;
subgraph cluster_test {
{
rank = same;
a0; b0; c0; d0; e0; f0;
}
{
rank = same;
c1; d1;
}
{
rank = same;
c2; d2;
}
}
}
The resulting graph is as follows:
What I want is for the ordering of level 0 nodes to be maintained, i.e, I want a0, b0 to come before c0, d0 in the horizontal direction.
How do I achieve this?
Empty nodes, edges with weight and explicit ordering of the top row in the cluster helps. See code below with annotations:
digraph so
{
// we define all nodes in the beginning, before edges and clusters
// may not be essential but I think it's good practice
start
a0 b0 c0 d0 e0 f0
c1 d1
c2 d2
end
// we define "empty" nodes that can be used to route the edges
node[ shape = point, height = 0 ];
ax bx ex fx
subgraph cluster_test
{
// we need to keep explicit order of the top nodes in the cluster
{ rank = same; a0 -> b0 -> c0 -> d0 -> e0 -> f0[ style = invis ] }
// the original layout in the cluster, empty nodes added at the bottom
{ rank = same; c1 d1 }
{ rank = same; ax bx c2 d2 ex fx }
c0 -> c1 -> c2;
d0 -> d1 -> d2;
// routing through invisible nodes keeps the position of all other nodes
// edges with no arrowheads, strong weight to keep it vertical
edge[ dir = none, weight = 10 ]
a0 -> ax;
b0 -> bx;
e0 -> ex;
f0 -> fx;
}
// connecting to the start and end node, normal edges again
edge[ weight = 1, dir = forw ];
start -> { a0 b0 c0 d0 e0 f0 }
{ ax bx c2 d2 ex fx } -> end;
}
which gives you
Related
How do I align nodes with each other between subgraphs?
In the example, I would like all the 0 nodes aligned vertically as well as the 1, 2, 3 and 4 nodes. It would be good if all the right edges of subgraphs were extended as necessary to align too.
digraph AlignWhenMissing {
rankdir=LR;
node [shape=box]
subgraph clusterA {label = "A Shift a4 along (no a3)";
a0 -> a1;
a1 -> a2;
a2 -> a4;
}
subgraph clusterB {label = "B Shift b2 above a2";
b0 -> b2;
b2 -> b3;
}
subgraph clusterC {label = "C Shift c3 above b3";
c0 -> c3;
}
A -> a0[lhead=clusterA];
B -> b0[lhead=clusterB];
C -> c0[lhead=clusterC];
a0 -> b0[constraint=false];
b3 -> c3[constraint=false];
a2 -> b2[constraint=false];
}
The minlen attribute (https://graphviz.org/docs/attrs/minlen/) applies to edges to lengthen them (i.e. to push them "out" by N ranks).
Then added invisible nodes and edges to increase cluster dimension.
digraph AlignWhenMissing {
rankdir=LR;
node [shape=box]
subgraph clusterA {label = "A Shift a4 along (no a3)";
a0 -> a1;
a1 -> a2;
a2 -> a4 [minlen=2];
}
subgraph clusterB {label = "B Shift b2 above a2";
b0 -> b2 [minlen=2]
b2 -> b3;
b4 [style=invis]
b3 -> b4 [style=invis]
}
subgraph clusterC {label = "C Shift c3 above b3";
c0 -> c3 [minlen=3];
c4 [style=invis]
c3 -> c4 [style=invis]
}
A -> a0[lhead=clusterA];
B -> b0[lhead=clusterB];
C -> c0[lhead=clusterC];
a0 -> b0[constraint=false];
b3 -> c3[constraint=false];
a2 -> b2[constraint=false];
}
Giving:
I am trying to make a graph where {C1;C2} are aligned vertically. Same with {A;B;C} and {E1;E2;E3}
Here is my code so far:
digraph G{
rankdir="LR";
S -> C1 [label="150"];
S -> C2 [label="130"];
C1 -> A [label="70"];
C1 -> B [label="80"];
C2 -> B [label="60"];
C2 -> C [label="70"];
C -> B [label="10"];
A -> E1 [label="70"];
B -> E1 [label="80"];
B -> D [label="70"];
C -> E3 [label="60"];
D -> E1 [label="40"];
D -> E2 [label="0"];
D -> E3 [label="30"];
E1 -> t [label="190"];
E2 -> t [label="0"];
E3 -> t [label="90"];
}
Here is the result so far:
{ rank = same; } is your friend. I have added invisible edges between A, B and C to keep them tighter together, and reversed one of your connections in orde to keep A over B over C etc.
digraph G
{
rankdir="LR";
{ rank = same; C1 C2 }
{ rank = same; A -> B -> C[ style = invis ] }
{ rank = same; E1 E2 E3 }
S -> C1 [label="150"];
S -> C2 [label="130"];
C1 -> A [label="70"];
C1 -> B [label="80"];
C2 -> B [label="60"];
C2 -> C [label="70"];
B -> C [dir = back, label="10"]; // to keep B above C
A -> E1 [label="70"];
B -> E1 [label="80"];
B -> D [label="70"];
C -> E3 [label="60"];
D -> E1 [label="40"];
D -> E2 [label="0"];
D -> E3 [label="30"];
E1 -> t [label="190"];
E2 -> t [label="0"];
E3 -> t [label="90"];
}
yields
If I run graphviz on this digraph:
digraph G {
subgraph cluster_0 {
style=filled;
color=lightgrey;
node [style=filled,color=white];
a0; a1; a2; a3;
label = "sources";
}
subgraph cluster_1 {
style=filled;
color=lightgrey;
node [style=filled,color=white];
b0; b1; b2; b3;
label = "intermediaries";
}
a0 -> b0; a1 -> b0;
a0 -> b1; a1 -> b1;
a2 -> b2; b0 -> b2;
b1 -> b2; a3 -> b3;
b0 -> b3; b1 -> b3;
}
I get
with many edges intersecting the "intermediaries" label. How do I get graphviz to make edges avoid labels?
As a workaround:
digraph G {
subgraph cluster_0 {
style=filled;
color=lightgrey;
node [style=filled,color=white];
a0; a1; a2; a3;
label = "sources";
}
subgraph cluster_1 {
style=filled;
color=lightgrey;
node [style=filled,color=white];
nodelabel [label="intermediaries"; style=filled; color=lightgrey]
nodelabel -> b1 [style=invis];
nodelabel -> b0 [style=invis];
b0; b1; b2; b3;
}
a0 -> b0; a1 -> b0;
a0 -> b1; a1 -> b1;
a2 -> b2; b0 -> b2;
b1 -> b2; a3 -> b3;
b0 -> b3; b1 -> b3;
}
produce:
I have code to display two subgraphs:
graph {
rankdir=LR;
subgraph cluster01 {
label="t=0"
a0 [label="A"];
a1 [label="B"];
a2 [label="C"];
a5 [label="E"];
a0 -- a1;
a1 -- a2 ;
a2 -- a0;
};
subgraph cluster02
{
label="t=10"
b0 [label="A"];
b5 [label="E"];
b1 [label="B"];
b2 [label="C"];
b0 -- b1;
b2 -- b5;
};
a0--b0 [style=dotted];
a1--b1 [style=dotted];
a2--b2 [style=dotted];
a5--b5 [style=dotted];
}
This code displays two subgraphs like this:
But I want to have it like this:
I hope someone will help me fix the "rankdir" to get it done.
The following was achieved by using invisible edges and constraint=false on some edges:
graph {
rankdir=LR;
subgraph cluster01 {
label="t=0";
a0 [label="A"];
a1 [label="B"];
a2 [label="C"];
a5 [label="E"];
a0 -- a1;
a1 -- a2;
a2 -- a5 [style=invis];
a2 -- a0 [constraint=false];
};
subgraph cluster02
{
label="t=10"
b0 [label="A"];
b5 [label="E"];
b1 [label="B"];
b2 [label="C"];
b0 -- b1;
b1 -- b2 [style=invis];
b2 -- b5;
};
edge[constraint=false];
a0--b0 [style=dotted];
a1--b1 [style=dotted];
a2--b2 [style=dotted];
a5--b5 [style=dotted];
}
I'm trying to use graphviz on media wiki as a documentation tool for software.
First, I documented some class relationships which worked well. Everything was ranked vertically as expected.
But, then, some of our modules are dlls, which I wanted to seperate into a box. When I added the nodes to a cluster, they got edged, but clusters seem to have a LR ranking rule. Or being added to a cluster broke the TB ranking of the nodes as the cluster now appears on the side of the graph.
This graph represents what I am trying to do: at the moment, cluster1 and cluster2 appear to the right of cluster0.
I want/need them to appear below.
<graphviz>
digraph d {
subgraph cluster0 {
A -> {B1 B2}
B2 -> {C1 C2 C3}
C1 -> D;
}
subgraph cluster1 {
C2 -> dll1_A;
dll1_A -> B1;
}
subgraph cluster2 {
C3 -> dll2_A;
}
dll1_A -> dll2_A;
}
</graphviz>
The layout is an attempt by Dot to minimise the overall height.
One reason for the more compact than required layout is the use of the edge that goes in the reverse direction from dll1_a to B1. It tries to pull the cluster as close back to the destination node as possible. To avoid this edge affecting the graph, either relax the constraint on the upwards edges as shown, or draw the edge in the forward direction and use the dir attribute to reverse the arrow.
This will help with many layouts but it alone is not sufficient to fix the example given. To prevent Dot from maintaining the compact layout it prefers you can add a minlen attribute to edges that should remain (near) vertical. This may be difficult to calculate in general but is practical for manually tuned layouts.
digraph d {
subgraph cluster0 {
A -> {B1 B2}
B2 -> {C1 C2 C3}
C1 -> D;
}
subgraph cluster1 {
C2 -> dll1_A [minlen = 2];
dll1_A -> B1 [constraint = false];
/* B1 -> dll1_A [dir = back]; */
}
subgraph cluster2 {
C3 -> dll2_A;
}
dll1_A -> dll2_A;
}
My experience shows that constraint=false commonly gives unnecessarily convoluted edges. It seems that weight=0 gives better results:
digraph d {
subgraph cluster0 {
A -> {B1 B2}
B2 -> {C1 C2 C3}
C1 -> D;
}
subgraph cluster1 {
C2 -> dll1_A [minlen = 2];
dll1_A -> B1 [weight = 0];
/* B1 -> dll1_A [dir = back]; */
}
subgraph cluster2 {
C3 -> dll2_A;
}
dll1_A -> dll2_A;
}
This will produce the graph you are looking for:
digraph d {
subgraph cluster0 {
A -> {B1 B2}
B2 -> {C1 C2 C3}
C1 -> D;
}
subgraph {
rankdir="TB"
subgraph cluster1 {
C2 -> dll1_A;
dll1_A -> B1;
}
subgraph cluster2 {
C3 -> dll2_A;
}
}
dll1_A -> dll2_A;
}
What this does is creat a subgraph that is used only for layout purposes to provide the top to bottom ordering that you desire.