If I am interpreting this correctly, dot method is calculating the scale of edges and nodes to be used. It is expecting at least one of the nodes to have n higher than 1, but if all nodes have n=1, then their maximum is. Logarithm of 1 is 0, and this is the denominator used for scaling, which ends up in a division by zero.
I added a try/except for this division and then the scale is just max_width - 1.
If I am interpreting this correctly, dot method is calculating the scale of edges and nodes to be used. It is expecting at least one of the nodes to have n higher than 1, but if all nodes have n=1, then their maximum is. Logarithm of 1 is 0, and this is the denominator used for scaling, which ends up in a division by zero.
I added a try/except for this division and then the scale is just max_width - 1.
This should solve issue #66 . Is this ok?