Closed jdramsey closed 7 years ago
@yuanzhou @espinoj
Maybe there's a better interface for Mimbuild, but this one at least works. Here's a screenshot.
Also, @ekummerfeld @ps7z
@jdramsey Do I just uncomment the config section and test things out or is there anything specific you want me to look at?
@yuanzhou I think just uncomment it and try it out, as you say. It should just work; it was working before, and it worked for me just now. I thought it better if you did the honors though, so you weren't surprised.
@jdramsey no problem. Since I'm also fixing another Tetrad bug, I may combine this issue and the one I'm working on in the same pull request later.
@yuanzhou Great, thanks!
I was able to get the Mimbuild work. Also noticed the weird highlighting issue that we noticed int he knowledge tier box.
@jdramsey for some reason, I could only get it work once. I tried several times later with multiple searches but just couldn't reproduce. What is the proper procedure to use the Mimbuild?
For example, I created a data model and imported some data. Then PC search1, FGES search2. On search3, I was able to select "Mimbuild" but it pops up this:
I think I'll need some instructions on how to use Mimbuild.
Try simulation first. In the Graph box make a random MIM, with four measures per latent, like this:
Do a search, as follows:
Then a measurement model box:
Then add Mimbuild:
This gives you a graph over the latents.
This is a little difficult to follow for a user; we may need to add an additional pipeline or two for it. Unless you can think of a better way to do it!
Thanks @jdramsey! I also set the cluster selection to start at 3 with a upper limit of 100. It was showing 0 even though we have 3 clusters, and the upper limit was 1000, which I think is too big.
I'll send a pull request soon.
I'm closing this since we've had Mimbuild restored in development
branch now.
In the config file, this bit of code needs to be uncommented: