omxGraphviz error

the code below generates and estimates the variance and mean of one variable, and should make a diagram like this:
http://openmx.psyc.virginia.edu/docs/OpenMx/latest/TwoModelStylesTwoDataStyles.html#model-specification

instead it generates an error

testData = as.matrix(rnorm (1000, 0, 1))
selVars c("X")
dimnames(testData) <- list(NULL, selVars)
univSatModel1 <- mxModel("univSat1", manifestVars= selVars,
mxPath(from="X", arrows=2, free=T, values=1, lbound=.01, labels="var X"),
mxData(observed=testData, type="raw"),
type="RAM"
)
fit = mxRun(univSatModel1)
omxGraphviz(fit)
# Error: mxPath() call will generate 0 paths but you have specified 1 arrows with 'from' argument assigned to and 'to' argument assigned to

This has already been fixed in the subversion trunk.

thanks for fixing the error, mike,

The function is not drawing the means triangle, just the measured variable and variance (double-headed) arrow.

should be this:

not this

I guess too that the dot file should add the estimated values to the paths (something close to 1 on the var and zero on the triangle for the generated data)

Some comments. Is this bug report accurate: http://openmx.psyc.virginia.edu/issue/2009/10/incorrect-figure-documentation? Also I'm going to start a forum post on SEM diagrams to get discussion.

the comment that documentation the first diagram shouldn't have a triangle, if it is to match the treatment of the data) is correct, but also you can see from summary(fit) that the failing script is the one which estimates two parameters: variance and mean (M matrix auto-created when data are raw in RAM)

/trunk/models/failing/omx_graphvis_example.R
# summary(fit)
# name matrix row col Estimate Std.Error
# 1 var X S 1 1 0.99325690 0.03114318
# 2 M 1 1 0.01635556 0.02262332

Oh I see. Which arrows get drawn? See http://openmx.psyc.virginia.edu/thread/244

Automatically closed -- issue fixed for 2 weeks with no activity.

I would just note en-passant that the documentation remains out of synch with the scripts (no means versus means)

Yes, the following ticket has not been resolved: http://openmx.psyc.virginia.edu/issue/2009/10/incorrect-figure-documentation