edgebundle doesn't render plot when in loop in markdown - r

Question

I'm trying to create an automated report where I create a series of chord graphs using edgebundleR.
I have a function that does a bunch of stuff and has more or less this form:
plot_chords <- function(x,t,pos) {
...
stuff I do with the data
...
g <- graph.adjacency(mydata, mode="upper", weighted=TRUE, diag=FALSE)
return(edgebundle(g))
}
This function works properly if I don't use it inside a loop. It doesn't if it is in a loop like this:
```{r echo = FALSE,message=FALSE, warning = FALSE,results = "asis"}
for (c in unique(df$Group)) {
cat("\n\n## ",c," - Negative Correlations (min r=",t_neg," - only significative)\n\n")
plot_chords(subset(df, Group == c),0.5,0)
}
```
I found that in general, this doesn't work inside loops unless I use print:
for (c in unique(df$Group)) {
temp=df[df$Group == c,]
print(plot_chords(temp,0.5,0))
}
But print doesn't work in markdown.
How can I render the plot?
Thanks.

Answer 1

The edgebundle call returns an htmlwidget and works, as you noted, well when not in a loop. A solution to your situation would be use the for loop to generate several specific R code chunks in a temporary file and then evaluate that that temporary file as a child file in your primary .Rmd file.
For example, in a .Rmd file these two chunks will load the needed packages and define a function foo which creates and shows a random edgebundle.
```{r}
set.seed(42)
library(edgebundleR)
library(igraph)
```
## test the function
```{r}
foo <- function() {
adjm <- matrix(sample(0:1, 100, replace = TRUE, prob = c(0.6, 0.4)), nc = 10)
g <- graph.adjacency(adjm)
edgebundle(g)
}
```
Calling foo twice in a chunk will work as expected in the output .html document.
```{r}
foo()
foo()
```
To generate several edgebudles in a for loop try this. Write a for loop to populate a temp.Rmd file with the needed R chunks. You will need to modify this as needed for your application.
## test the function in a for loop
```{r}
tmpfile <- tempfile(fileext = ".Rmd")
for(i in 1:3) {
cat("### This is edgebundle", i, "of 3.\n```{r}\nfoo()\n```\n",
file = tmpfile, append = TRUE)
}
```
The contents of tmpfile look like this:
### This is edgebundle 1 of 3.
```{r}
foo()
```
### This is edgebundle 2 of 3.
```{r}
foo()
```
### This is edgebundle 3 of 3.
```{r}
foo()
```
To display the widgets in your primary output file use a chunk like this:
```{r child = tmpfile}
```
The full .Rmd file and result:
example.Rmd:
# edgebundleR and knitr
Answer to https://stackoverflow.com/questions/47926520/edgebundle-doesnt-render-plot-when-in-loop-in-markdown
```{r}
set.seed(42)
library(edgebundleR)
library(igraph)
```
## test the function
```{r}
foo <- function() {
adjm <- matrix(sample(0:1, 100, replace = TRUE, prob = c(0.6, 0.4)), nc = 10)
g <- graph.adjacency(adjm)
edgebundle(g)
}
foo()
foo()
```
## test the function in a for loop
```{r}
tmpfile <- tempfile(fileext = ".Rmd")
for(i in 1:3) {
cat("### This is edgebundle", i, "of 3.\n```{r}\nfoo()\n```\n",
file = tmpfile, append = TRUE)
}
```
```{r child = tmpfile}
```
```{r}
print(sessionInfo(), local = FALSE)
```