Page 1 of 2
scaling problem...
Posted: Tue Nov 06, 2012 1:04 pm
by bunker
Working with latest buil I encountered lot of trouble doing "simple" things like scaling on model I used in the past without problem.
I converted my old model into new one but I notice this problems are present with models actually in latest build.
For example after you have done one scaling you cannot change any parameters like in the past and improve precision, for example changing scale factor or activate/deactivate option. After first run the "RUN" button is greied out and you need to close and reopen the scale pane to do another scaling...
This is very annoying...but is not the only problem...
In fact after first run of scaling I try to deactivate "Preview static pose" and every time I receive an error (scaling is still running) and I need to close Opensim to stop it ("x" button not stop it).
Here you can se an image. The model was prviously scaled but after I have deactivated the option and relaunch the scaling...
[img]
[/img]
Thank you a lot for your help.
Andrea
Re: scaling problem...
Posted: Wed Nov 07, 2012 12:51 am
by bunker
This is another example...
After 1st scaling I deactivate the "Preview static pose" mark... RUN and have had the error.
Andrea
[img]
[/img]
Re: scaling problem...
Posted: Wed Nov 07, 2012 5:03 am
by bunker
this is the java error...
java.lang.RuntimeException: SimTK Exception thrown at Xml.cpp:77:
Error detected by Simbody method Xml::writeToFile(): Failed to write to the Xml file 'C:\MyOpenSimModel\Davis\bici\C:\MyOpenSimModel\Davis\bici\Unassigned' with error ' (line=1, col=1)'.
(Required condition 'saveOK' was not met.)
at org.opensim.modeling.opensimModelJNI.OpenSimContext_processModelMarkerPlacer__SWIG_0(Native Method)
at org.opensim.modeling.OpenSimContext.processModelMarkerPlacer(OpenSimContext.java:321)
at org.opensim.tracking.ScaleToolModel$ScaleToolWorker.construct(ScaleToolModel.java:343)
at org.opensim.swingui.SwingWorker$2.run(SwingWorker.java:108)
[catch] at java.lang.Thread.run(Thread.java:662)
Re: scaling problem...
Posted: Wed Nov 07, 2012 7:55 am
by bunker
update....
I find a trick... after first scale if you close the scale panel and do another scaling from scratch on the model the error is not there...
Andrea
Re: scaling problem...
Posted: Wed Nov 07, 2012 10:59 am
by aymanh
Hi Andrea,
Thanks for reporting. I was able to reproduce your problem (it's caused by mangling the name of the output model file when you run the tool more than once). As you found out the workaround for now is to close the scale tool, change current model to the original model and then relaunch the scale tool.
Please file a bug report so that you can track progress and verify the fix in the next public release.
Best regards,
-Ayman
Re: scaling problem...
Posted: Fri Nov 09, 2012 3:18 pm
by bunker
OK, thank you Aymanh I will.
Andrea
Re: scaling problem...
Posted: Wed Nov 14, 2012 8:30 am
by francescorea1982
Hi everybody,
This problem happens to me as well. It seems that scaling (even the first time) any model causes the error.
Unfortunately the workaround does not seem to work for me.
Any other idea on how I can bypass the error.
Many Thanks
Francesco
Re: scaling problem...
Posted: Wed Nov 14, 2012 10:09 am
by aymanh
Hi Francesco,
Since the example files included in the distribution can run Scale (at least for the first time) without problem, you can compare your setup file with the file provided with the distribution for gait2354 and remove the intermediate file specifications (for MarkerSet, ScaleSet, etc.) so that Scale runs to completion.
Another option is to run the scale tool from the command line on the setup file (since the issue has to do with GUI handling of file paths) and then open the output model in the GUI.
This's a temporary workaround until we post an update that fixes the problem more robustly.
Best regards,
-Ayman
Re: scaling problem...
Posted: Wed Jan 30, 2013 1:48 pm
by buckeyes4ever
Hello Ayman,
I have recently switched to 3.0 and am also having trouble while scaling a model set up for our lab. I have been able to scale the tutorial file and Hamner’s running model in 3.0, and am confused as to why our model won’t scale. (It is set up very similarly to Hamner’s model.)
The error occurs when I run the scaling tool; it gets through “step 3” in the scripting shell window, and then times out after stating “deleted 0 unused markers from model”. A bar continuously updates in the bottom right that says “executing scaling”, but it never finishes.
I tried the first work-around mentioned on this thread (scaling another model, and then trying) but that didn’t work. I also tried to scale the model using the command prompt (scale –S <file name.xml>) but that had its own problems, namely continuously referring to a file named “unassigned”:
Step 1: Loading Generic Model
Loaded model ScaleSet from file Unassigned
Step 2: Scaling Generic Model
wrote model file Unassigned from model {file name}
wrote scale file Unassigned for model {file name}
Step 3: Placing markers on model
Loaded marker file {file name.trc} <82 markers, 45 frames>
Averaged frames from time 0 to .14667 in {file name.trc} <frames 1 to 45>
Deleted 0 unused markers from model {file name}
InverseKinematicsSolver: Model has no markers!
Exception:
InverseKinematicsSolver: Model has no markers!
Two things jumped out at me when looking at this:
1. I used the command prompt to scale the tutorial file, and during step 1 it referenced the .osim file, whereas mine references “unassigned”. Is there a reason the command does not know what my .osim file is?
2. The error occurs because “Model has no markers” yet 3 lines before that it says the model has 82 markers. Why the discrepancy?
Myself and a few lab mates have been trouble shooting this for a while and feel pretty stuck. We’d like to use 3.0 going forward, but may have to revert back to 2.4 if we cannot get past this.
Thank you so much for any help you could give!
Justin
Re: scaling problem...
Posted: Thu Jan 31, 2013 2:02 pm
by aymanh
Hi Justin,
Please try the beta version of 3.0.1 (already posted) and let me know if the problem has been fixed there.
Thanks much,
-Ayman