The nodes of tree are overlap when plot a tree. Maybe a bug when plot compex trees.

[yizhichun]

• This issue is for the Python interface of igraph.
• This issue is a bug report or a feature request, not a support question.

The nodes of tree are overlap when plot a complex tree, as shown in attached figure.
May be has a bug.
My tree structure as following:

edges = 
[(0, 1), (1, 2), (2, 3), (3, 4), (3, 5), (2, 6), (6, 7), (6, 8), (8, 9), (9, 10), (8, 11), (11, 12), (8, 13), (6, 14), (14, 15), (14, 16), (14, 17), (14, 18), (1, 19), (19, 20), (20, 21), (20, 22), (22, 23), (22, 24), (22, 25), (20, 26), (26, 27), (27, 28), (27, 29), (27, 30), (26, 31), (26, 32), (32, 33), (32, 34), (26, 35), (35, 36), (35, 37), (1, 38), (38, 39), (38, 40), (38, 41), (41, 42), (41, 43), (1, 44), (44, 45), (45, 46), (45, 47), (47, 48), (48, 49), (47, 50), (50, 51), (47, 52), (52, 53), (44, 54), (54, 55), (54, 56), (56, 57), (57, 58), (58, 59), (58, 60), (57, 61), (56, 62), (56, 63), (63, 64), (44, 65), (65, 66), (66, 67), (66, 68), (68, 69), (65, 70), (65, 71), (71, 72), (72, 73), (65, 74), (65, 75), (75, 76), (75, 77)]

The plot codes as following:

from igraph import *
tree = Graph(edges = edges)
plot(tree,layout = tree.layout_reingold_tilford(root=[0]))

How to fix this, thank you!

[ntamas]

Yes, this is a bug in the underlying implementation of the Reingold-Tilford algorithm in the C core of igraph. I don't have too much spare time to deal with this, so if you would like to help, can you please try to construct a minimal example that still exhibits the problem, e.g., by removing leaf nodes one by one from the tree until the issue does not appear any more? This would be a great help to me so I could start exploring with the smallest example that is possible to be derived from the current one.

[yizhichun]

I am glad to help. This will be much helpful for plot-ing complex trees.

After testing a lot, I have found a typical example that exhibits the problem. As following:

from igraph import *
edges = [(0, 1),
 (1, 2),
 (2, 3),
 (2, 4),
 (2, 5),
 (2, 6),
 (0, 7),
 (7, 8),
 (7, 9),
 (7, 10),
 (10, 11),
 (10, 12),
 (10, 13),
 (10, 14)]
tree = Graph(edges=edges)
tree.vs['label'] = [v.index for v in tree.vs]
plot(tree,layout=tree.layout('rt',root=[0]),bbox=(300,300))

The keys of this example are like this: the nodes in Level 4 will be out of the drawing limitation, when there are two nodes, 8 and 9, on the left of 10, in the Level 3.

If the four nodes in Level 4 are children of node 9, which means just one node 8 on the left of 9, the trees drawing well.

If there are two node on the right of left root in Level 3, the trees are also drawing well.

If the leaves are in Level 3, the trees are also drawing well.

If the leaves are in Level 4 or higher, this bug will be appear.

I hope this would be helpful. I'm pleasure to do whatever else I can do to fix it.
Thanks.

[yizhichun]

Another example is like this:

from igraph import *
edges = [(0, 1),
 (1, 2),
 (2, 3),
 (0, 4),
 (4, 5),
 (4, 6),
 (0, 7),
 (7, 8),
 (8, 9),
 (8, 10),
 (8, 11),
 (8, 12),
 (8, 13),
 (8, 14),
 (8, 15),
 (8, 16)]
tree = Graph(edges=edges)
tree.vs['label'] = [v.index for v in tree.vs]
plot(tree,layout=tree.layout('rt',root=[0]),bbox=(300,300))

[doRemy]

Thank you for your help to understand this issue.

Without modifying the edges list, how would you go about finding a solution to this problem?

Is there a layout option to prevent these kind of overlapping nodes?

[szhorvat]

Likely the same as igraph/igraph#879

[ntamas]

Yes, it's the same issue. Let's continue the discussion there as it has to be fixed in the C layer.