Re: [igraph] Turn a directed network into a weighted undirected network

[MikeS]

Hello, all

Tamas, thanks you for ideas.

I found usefull links
http://www.mikedecuir.com/archive/2012/12/finding-triangles-with-r/
https://graph-tool.skewed.de/static/doc/topology.html#graph_tool.topology.subgraph_isomorphism

I have tried to implicate the your approach.
Unfortunately, I couldn't to map the vertex IDs to edge IDs using
get.edge.ids() on step 2. I have tried

ei <- get.edge.ids(UG, c(1,2, 2,3, 3,1))
E(UG)[ei]

This code works, but I couldn't understand how to use a list in common
case. Without step 2, I couldn't to code step 4. Here's my current
attempt.

library(igraph)

d <- matrix(c(0,1,0,1,0, 0,0,1,1,1, 1,0,0,1,1, 0,1,1,0,1,
1,1,1,0,0),nrow=5,ncol=5)
G <- graph.adjacency(d)
V(G)$label <-c("A","B","C","D","E")
V(G)$shape = "none"
plot(G, edge.arrow.size=0.5)

# Here are the adjacency matrices for each of the four subgraphs

d0<-matrix(c(0,1,0,0,0,1,1,0,0),nrow=3,ncol=3)
d1<-matrix(c(0,1,0,0,0,1,1,1,0),nrow=3,ncol=3)
d2<-matrix(c(0,1,0,1,0,1,1,1,0),nrow=3,ncol=3)
d3<-matrix(c(0,1,1,1,0,1,1,1,0),nrow=3,ncol=3)

# Turn them into a convenient list
sbgCycle.mat<-list(d0,d1,d2,d3)
n <- length(list(d0,d1,d2,d3))

# And then into a list of graph objects

pattern<-lapply(sbgCycle.mat, graph.adjacency)

par(mfrow=c(2,2))

# The four triangles all contain a directed 3-cycle subgraph
plot(pattern[[1]], edge.curved=TRUE, xlab='Cycle')
plot(pattern[[2]], edge.curved=TRUE, xlab='One reciprocal edge')
plot(pattern[[3]], edge.curved=TRUE, xlab='Two reciprocal edges')
plot(pattern[[4]], edge.curved=TRUE, xlab='Three reciprocal edges')

# 1. Convert the graph to undirected
UG <- simplify(G)
UG <- as.undirected(UG)
#plot(UG)

# 2. Search for triangles in the undirected graph

triangle     <- graph.full(3)
sbg.triangle <- graph.get.subisomorphisms.vf2(UG, triangle)
ei <- get.edge.ids(UG, c(1,2, 2,3, 3,1)) # work
E(UG)[ei]
#e1 <- get.edge.ids(UG, c(sbg.triangle[[1]])) # does't work


# 3. Search for subisomorphisms in the directed graph to all of the
four templates

#for(i in 1:n){ # loop 'for' does not work ? rewrite with 'apply' later
subisom1 <- subgraph_isomorphisms(pattern[[1]], G, method="lad", induced=TRUE)
subisom2 <- subgraph_isomorphisms(pattern[[2]], G, method="lad", induced=TRUE)
subisom3 <- subgraph_isomorphisms(pattern[[3]], G, method="lad", induced=TRUE)
subisom4 <- subgraph_isomorphisms(pattern[[4]], G, method="lad", induced=TRUE)
#}

# 4. For each triangle check which pattern it was isomorphic

G2 <- G
for(i in 1:length(sbg.triangle))
{
  new_weight <- subgraph_isomorphisms(pattern[[1]], G, method="lad",
induced=TRUE)
  E(G2, path = E(G)[get.edge.ids(???)])$weight = new_weight
}

--
Mike