Include parameter estimates on .Dot plots from omxGraphviz()

In terms of SEM competition under R, both lavaan[sp] and sem 2.0 have nice reporting of graphical output with parameters, as well as comprehensive fit indices (inc. for example saturated fit for RAM models)

We have most of this now, and worth going extra metre or two to complete it.

getting omxGraphviz to include values on the path diagram is pretty straight forward.

An initial requirement given the current implementation is to modify
OpenMx:::matrixToPaths() and meansToPaths to include values

something like this would work

matrixToPaths <- function (mxMatrix, arrows = c(1, 2)) {
values <- mxMatrix@values
free <- mxMatrix@free
labels <- mxMatrix@labels
if (arrows == 2) {
values[upper.tri(values)] <- 0
free[upper.tri(free)] <- FALSE
labels[upper.tri(labels)] <- as.character(NA)
}
select <- (values != 0) | (free) | (!is.na(labels))
if (length(select) > 0) {
rowFactors <- row(values, as.factor = TRUE)
colFactors <- col(values, as.factor = TRUE)
fromNames <- as.character(colFactors[select])
toNames <- as.character(rowFactors[select])
pathValues <- values[select]

# return(pathValues)
if (length(fromNames) > 0 && length(toNames) > 0) {
return(mxPath(from = fromNames, to = toNames, arrows = arrows, values= pathValues))
}
}
return(list())
}

meansToPaths <- function (mxMatrix) {
if (is.null(mxMatrix))
return(list())
values <- mxMatrix@values
free <- mxMatrix@free
labels <- mxMatrix@labels
select <- (values != 0) | (free) | (!is.na(labels))
if (length(select) > 0) {
colFactors <- col(values, as.factor = TRUE)
toNames <- as.character(colFactors[select])
pathValues <- values[select]

if (length(toNames) > 0) {
return(mxPath(from = "one", to = toNames, arrows = 1, values=))
}
}
return(list())
}

Then modifying writeDotFile() to include "label = $value" in the paths as it creates them

happy to help code this

t