[BioC] Rgraphviz produces segmentation fault when attempting to plot Bayesian Belief Networks

[Royce Francis]

Hello all,

I have been studying the use of Bayesian Belief Networks, and have found Marco Scutari's package, "bnlearn," very helpful.  While I was learning his package, I had been using graphviz to plot the networks that were learned.

Recently, I have not been able to plot any networks using graphviz because each time I call the function, graphviz.plot, I receive a message indicating that a segmentation fault has occurred.  I originally thought this was caused by updating R to 2.14.0, but when I reverted to 2.13.0, the segmentation faults persisted.

Here is a snippet of code using an example in Scutari's JSS paper that will produce the error in my setup (and the corresponding output):

> library("bnlearn")
> library(Rgraphviz)
> data(learning.test)
> bn.hc <- hc(learning.test,score="aic")
> graphviz.plot(bn.hc)

 *** caught segfault ***
address 0x10, cause 'memory not mapped'

Traceback:
 1: .Call("Rgraphviz_doLayout", graph, as.integer(type), PACKAGE = "Rgraphviz")
 2: graphLayout(g)
 3: agopen(x, name = name, layoutType = layoutType, recipEdges = recipEdges,     nodeAttrs = nodeAttrs, edgeAttrs = edgeAttrs, attrs = attrs,     ...)
 4: layoutFun(x, ...)
 5: layoutGraph(graph.obj, attrs = attrs, nodeAttrs = node.attrs,     layoutType = layout)
 6: graphviz.backend(nodes = nodes, arcs = arcs, highlight = highlight,     layout = layout, shape = shape, main = main, sub = sub)
 7: graphviz.plot(bn.hc)

Possible actions:
1: abort (with core dump, if enabled)
2: normal R exit
3: exit R without saving workspace
4: exit R saving workspace
Selection: 

I would greatly appreciate any help resolving this error, or any suggestions for alternative plotting methods.  I find the "plot" function completely inadequate and would much rather use graphviz to report my findings.  

Thank you in advance,
Royce