(0028741)
|
Greg Weirs
|
2012-06-29 16:15
|
|
I had an issue today when displaying two different sesame files in the same view, in which the auto scaling was not applied in the same way to both files. I turned on the cube axes for each surface, and they did not match - a given value of the pressure for one surface was not at the same height as the same value for the other surface. Similar observations for all three axes. Log scaling is not the culprit, log scaling was never turned on for any of the axes for either sesame file. Also I did not modify the prism view control at all. I know this bug is for the prism view control, but I think it is the same bug exposed by a different mechanism.
Also, I observed this while working with a user, so there is someone waiting for a fix. |
|
(0029125)
|
Utkarsh Ayachit
|
2012-09-04 11:50
|
|
Greg,
Can give me exact steps to reproduce this please? I don't look at Prism often enough to even remember exactly how to show multiple surfaces in the same view, etc. If you (or Alan) could give me exact steps using the data files we have in ParaViewData (or any other ones that you can send me), that would be a huge help.
For example,
1> Open PrismBlock.e
2> Click "Create Prism View"
3> "Apply"
4> Close "3D View"
.... |
|
(0029596)
|
Greg Weirs
|
2012-10-30 13:02
|
|
I have found one repeatable case. Pay attention to where the APPLY button is pressed. Steps:
1.) open "SymmetricImpact3.exo"
2.) select all variables to load and APPLY
3.) load the Prism Plugin (Tools->Manage Plugins)
4.) Add prism filter to the pipeline. Of the three prism plugin icons on the toolbar, it is the leftmost prism splitting light. Sometimes it is greyed out at this point (that's a BUG, but not the main one,) if so unhighlight or unselect the exodus reader in the pipeline viewer and then re-highlight it.
A dialog box opens and wants you to pick a file with the SESAME data. Pick "proxium.asc"
If successful, a second view opens.
5.) prism filter properties:
pick warp surface variable as "Total EOS (Pressure)" (Warp surface section)
select Display Contours checkbox (Display Contours section)
Select temperature as the contour variable
select "new range" and accept defaults
pick "SESAME to SI" units checkbox (Conversions section)
At the top of the properties panel there are two tabs, "SESAME Data" that you have been working with and "Simulation Data". Now pick the "Simulation Data." Associate the x, y, and z axes with DENSITY1, TEMPERATURE1, and PRESSURE1, respectively
back to SESAME Data tab
5.) Press APPLY
In the second view you will probably just see a single small square, i.e. a Points representation of a single data value, as opposed to surface, wireframe, etc.) If so, press the Camera reset button. Then you should see a large grey almost square surface, with high z values at the right (high x). (That you have to press the camera reset button is a BUG, but again, not the main one.)
If you now press the third prism plugin icon ("Change Prism View Scale"), which looks like a reverse of the first and with four red squares in the corners, a dialog window opens. There are three columns and three rows, with valid data for the Full Bounds and Threshold Bounds for each axis. Valid ranges: x Axis 20-3225, y Axis 105-800, z Axis 168619-6.39537e9. Custom scale will be greyed out and with ranges 0-1 for all axes. Click Cancel (Pop-up closes).
Main BUG behavior.
On the Properties Panel, check one of the boxes in the Curve Tables section. ("Show Table 40X...")
Press APPLY
Now click the third prism plugin icon again. For Full Bounds and Threshold Bounds, all ranges will be 1 to -1 ("1 - -1"). None of these are valid. At this point, there are no problems with the data displayed in the view, but when the invalid ranges are present in the Change Prism View Scale dialog, Prism's automatic scaling of the axes is in a broken state. The user cannot use the dialog to rescale the axes, and if more than one SESAME data set is in the prism view, the different data sets will not be scaled consistently. So, invalid ranges in the bounds is the main way I can see that something is wrong.
If you uncheck all the Curve Tables and APPLY, the valid bounds are recovered. I see invalid bounds when the Display Contours box is checked and at least one of the Curve tables boxes is checked. if just curve tables, or just Display Contours, I see valid bounds.
This is one way to see the BUG, but I think there may be more than one place in the software to fix. The initial reports were with sesame data that did not have any curve tables, so both the curve tables and contours were never on. But the bounds were still invalid. Once the bounds are invalid, I have never seen the plugin recover correct bounds until I found the steps above. So, to me, this suggests that there is more than one path to the BUG behavior. |
|