ParaView issueshttps://gitlab.kitware.com/paraview/paraview/-/issues2024-03-07T20:47:47-05:00https://gitlab.kitware.com/paraview/paraview/-/issues/18060Volume rendered clip planes are bad2024-03-07T20:47:47-05:00W. Alan ScottVolume rendered clip planes are badHere is a state file, created by Dave K, showing how bounding box clip planes are incorrect for volume rendering. Uses wavelet. Note that it is creating a very long, skinny wavelet, then clipping, then using the transformation scale, i...Here is a state file, created by Dave K, showing how bounding box clip planes are incorrect for volume rendering. Uses wavelet. Note that it is creating a very long, skinny wavelet, then clipping, then using the transformation scale, in the Z axis, to rescale to a cube.
* Linux, 5.5.0-rc3, local server.
* Load state file: broken-projected-tetrahedra.pvsm.
You will see nothing.
* -Y.
You now see what should be visible.
* Left mouse button, drag down maybe 10 degrees. You now see the cube disappearing from the bottom.
* Drag a bit more. More of the cube disappears.
* To see the cube totally disappeared, +Z.
* To see the cube again, -Y.
[broken-projected-tetrahedra.pvsm](/uploads/f062e87e3a62a8242c930dd0f896ea37/broken-projected-tetrahedra.pvsm)5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/18058No Valgrind/Test Coverage in ParaView's Cdash2020-12-10T21:29:07-05:00Jonathan BorduasNo Valgrind/Test Coverage in ParaView's CdashHi all,
The vtkDebugLeaks is fun, but quite limited in comparison with valgrind informations.
Moreover, the lack of test coverage in the ParaView's CDash makes it hard to know what is well tested, what might be deprecated and go unnot...Hi all,
The vtkDebugLeaks is fun, but quite limited in comparison with valgrind informations.
Moreover, the lack of test coverage in the ParaView's CDash makes it hard to know what is well tested, what might be deprecated and go unnoticed by the community using those functions.
Best regards,
Jonathan Borduashttps://gitlab.kitware.com/paraview/paraview/-/issues/18052axis grid text size is too small/ dim2024-03-07T20:47:47-05:00W. Alan Scottaxis grid text size is too small/ dimPlease change Axis Grid and Data Axis Grid defaults to be as follows:
5.5.0-rc3, Linux.
* X,Y,Z Title Font Properties. Font size 14, Bold.
* X,Y,Z Axis Label Font Properties. Font size 14, Bold.Please change Axis Grid and Data Axis Grid defaults to be as follows:
5.5.0-rc3, Linux.
* X,Y,Z Title Font Properties. Font size 14, Bold.
* X,Y,Z Axis Label Font Properties. Font size 14, Bold.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/18051Hidden line removal doesn't work with Ospray2023-06-15T09:16:02-04:00W. Alan ScottHidden line removal doesn't work with OsprayHidden line removal doesn't work with OSPRay.
* 5.5.0-rc3, Linux, remote server (shouldn't matter).
* Sources/ Unstructured Cell Type
* Wireframe.
* Hidden line Removal.
* Enable OSPRay.
The back lines should be removed.Hidden line removal doesn't work with OSPRay.
* 5.5.0-rc3, Linux, remote server (shouldn't matter).
* Sources/ Unstructured Cell Type
* Wireframe.
* Hidden line Removal.
* Enable OSPRay.
The back lines should be removed.https://gitlab.kitware.com/paraview/paraview/-/issues/18046incorrect legend label values and positions when switching on/off 'log scale'...2021-03-23T18:23:58-04:00Yanxiangincorrect legend label values and positions when switching on/off 'log scale' for variable containing zero__Version:__ 5.4.1 64-bit
__Platform:__ CentOS 7.3
__Issue:__ Text labels in the legend are not placed at the correct positions after switching on/off the log scale for a variable that has zero value.
__Steps to reproduce:__
1. show ...__Version:__ 5.4.1 64-bit
__Platform:__ CentOS 7.3
__Issue:__ Text labels in the legend are not placed at the correct positions after switching on/off the log scale for a variable that has zero value.
__Steps to reproduce:__
1. show the surface plot of a slice for a variable that contains value zero using 'log scale' (in the `Color Map Editor`). Warning message will show and ParaView automatically resets the lower bound of the range to 1.0
2. switch off 'log scale' and now the labels are incorrectly positioned, and the lower bound is always 0. However, the surface plot is rendered with the correct range. This is confirmed by adding the range annotation (see attached image below).
![legend](/uploads/f450d2e2650e272b83709ddaf69b7a5a/legend.png)
__NOTE:__
It was not intentional to switch on 'log scale' for variable that has 0 or negative values. This setting is automatically applied (randomly) when it is used for another variable (which could potentially be another bug).https://gitlab.kitware.com/paraview/paraview/-/issues/18045vtkPythonProgrammableFilter->SetParameter with multi line string does not work2021-03-23T18:25:26-04:00Shuhao WuvtkPythonProgrammableFilter->SetParameter with multi line string does not workTrying to set a multiline string via SetParameter for the programmable filter will not set the string correctly as it simply interpolates it into the script as oppose to escaping it first. This causes a syntax error.
Example of trying t...Trying to set a multiline string via SetParameter for the programmable filter will not set the string correctly as it simply interpolates it into the script as oppose to escaping it first. This causes a syntax error.
Example of trying to set `scalar_str` to `hello world\nhello world`:
```
SyntaxError: invalid syntax
Traceback (most recent call last):
File "<string>", line 22, in <module>
NameError: name 'RequestData' is not defined
File "<string>", line 6
scalar_str = r'hello world
^
SyntaxError: EOL while scanning string literal
Traceback (most recent call last):
File "<string>", line 22, in <module>
NameError: name 'RequestData' is not defined
```https://gitlab.kitware.com/paraview/paraview/-/issues/18043Box Chart: series order doesn't always match between the Panel and the View2021-03-23T18:25:39-04:00Utkarsh AyachitBox Chart: series order doesn't always match between the Panel and the View* Open [tapered_cyl.p3d](https://gitlab.kitware.com/berkgeveci/flow-vis-with-vtk/tree/master/tapered_cyl) dataset
* Apply "Compute Quartiles" filter
* Now click and drag the variables in **Series Parameter** widget to reorder them vertic...* Open [tapered_cyl.p3d](https://gitlab.kitware.com/berkgeveci/flow-vis-with-vtk/tree/master/tapered_cyl) dataset
* Apply "Compute Quartiles" filter
* Now click and drag the variables in **Series Parameter** widget to reorder them vertically. The view doesn't update. The view should have matched the order in the widget. Same is true if one reorders by clicking and dragging in the view itself.https://gitlab.kitware.com/paraview/paraview/-/issues/18029Displaying Points only data2022-06-09T14:09:54-04:00Robert M. O'BaraDisplaying Points only dataWhen dealing with vtkData that contains no cells (only points), have ParaView use a default representation (like Point Gaussian) that is more appropriate.When dealing with vtkData that contains no cells (only points), have ParaView use a default representation (like Point Gaussian) that is more appropriate.https://gitlab.kitware.com/paraview/paraview/-/issues/18025Add command line argument section to Users Guide2024-03-07T20:47:47-05:00W. Alan ScottAdd command line argument section to Users GuidePlease add a dedicated section in the Users Guide to command line arguments, for both paraview, pvserver, pvbatch and pvpython (assuming they are different). This is primarily driven by the need to explain stereo.Please add a dedicated section in the Users Guide to command line arguments, for both paraview, pvserver, pvbatch and pvpython (assuming they are different). This is primarily driven by the need to explain stereo.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/18019internal surfaces visible in certain parallel rendering scenarios2024-03-07T20:47:47-05:00Utkarsh Ayachitinternal surfaces visible in certain parallel rendering scenarios* Paraview + pvserver (at least 2 ranks), ensure that you're remote rendering by settings Remote Rendering Threshold to 0 in the settings dialog.
* Wavelet, Apply
* On Properties panel, change **Backface Representation** to **Cull Frontf...* Paraview + pvserver (at least 2 ranks), ensure that you're remote rendering by settings Remote Rendering Threshold to 0 in the settings dialog.
* Wavelet, Apply
* On Properties panel, change **Backface Representation** to **Cull Frontface**. Interact. Notice you see a nice hollow box as expected.
* Apply Clip, and change **Backface Representation** to **Cull Frontface**. This time, alas, inner faces are visible.
![image](/uploads/7676ebdb124afb8f5d88644f5a3d15fd/image.png)
Relates to #11811, #125465.13 (Summer 2024)Spiros TsalikisSpiros Tsalikishttps://gitlab.kitware.com/paraview/paraview/-/issues/18014Change plot anything over time to optionally merge all blocks2024-03-07T20:47:47-05:00W. Alan ScottChange plot anything over time to optionally merge all blocksOur current plot selection over time, and plot data over time filters both are very block based. Usually, what users want is not one or numerous blocks, but to see their data presented as a single result. In other words, make the optio...Our current plot selection over time, and plot data over time filters both are very block based. Usually, what users want is not one or numerous blocks, but to see their data presented as a single result. In other words, make the option (default on) to merge blocks before running these filters.
Well, Ken had a good point (as always). You probably want to run the filter as is, under the cover, and then combine the results of all blocks. This will be more efficient than merging blocks at the start.5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/18009static axis grid needed2024-03-27T16:23:29-04:00W. Alan Scottstatic axis grid neededWe need the ability to have an axes grid that does not move over time. The extends should default to time zero for the dataset, but could be custom set by the user. For instance, with can.exo, you could see where the original, pre-disp...We need the ability to have an axes grid that does not move over time. The extends should default to time zero for the dataset, but could be custom set by the user. For instance, with can.exo, you could see where the original, pre-displaced axes grid is. A trick that is being used currently is to put a sources/ box around the object, and then using Data Axes Grid. Lets make it simple for our users!5.13 (Summer 2024)https://gitlab.kitware.com/paraview/paraview/-/issues/18008Axes grid reset resets too much2018-02-26T21:13:07-05:00W. Alan ScottAxes grid reset resets too muchThe Axes Grid reset (at the bottom of the Axis Grid Edit panel) also turns the axis grid itself off. I think it should reset this whole panel, but then leave the Axes Grid in the state (on or off) that it was in.The Axes Grid reset (at the bottom of the Axis Grid Edit panel) also turns the axis grid itself off. I think it should reset this whole panel, but then leave the Axes Grid in the state (on or off) that it was in.https://gitlab.kitware.com/paraview/paraview/-/issues/18003Clip and Slice Incorrect for Unstructured Polyhedron Mesh2021-03-23T18:10:24-04:00Duncan KarnitzClip and Slice Incorrect for Unstructured Polyhedron MeshI have unstructured mesh data based on an Octree mesh. I have written it to files using the Ensight Gold NFaced unstructured mesh element types.
When I attempt to clip or slice the data set I see a few issues.
1) Slices often are missi...I have unstructured mesh data based on an Octree mesh. I have written it to files using the Ensight Gold NFaced unstructured mesh element types.
When I attempt to clip or slice the data set I see a few issues.
1) Slices often are missing whole cells in the data set.
2) Clipping often shows either elements protruding from the clipping plane or recessed from the clipping plane depending on the plane normal.
The missing cells in the slice and the protruding/recessed cells in the clip are the same.
I am quite sure there are no issues in the mesh connectivity. I can visualize the data set well using either a volume rendering or by first using the Tetrahedralize filter. However, even when using the tetrahedralize filter I can generate spherical slices that show missing pieces that would correspond to the missing cells when slicing the original data set.
I've attached images showing the general element edge lattice (Extract Edges) and the described slicing and clipping behavior. I've also attached the EnSight gold case/geo/data file set as a tgz archive.
[octree-sample.tgz](/uploads/ce828c20d3682296635be4d18733169e/octree-sample.tgz)
![octree-extract-edges](/uploads/c321f7dfc2f28999bebffe27554e7c2d/octree-extract-edges.png)
![octree-slice-missing-cells](/uploads/35474e3c4813c876a9527ab670dee155/octree-slice-missing-cells.png)
![octree-clip-protruding-cells](/uploads/39430b6babdc6b9d823f9266b59d37d9/octree-clip-protruding-cells.png)https://gitlab.kitware.com/paraview/paraview/-/issues/17997Python script UI needs line numbers2018-02-16T19:50:02-05:00Utkarsh AyachitPython script UI needs line numbersWhen typing script in Python Programmable Filter/Source, and if there are errors, line numbers are reported in error messages. However the UI has no line numbers which makes it hard to use.When typing script in Python Programmable Filter/Source, and if there are errors, line numbers are reported in error messages. However the UI has no line numbers which makes it hard to use.https://gitlab.kitware.com/paraview/paraview/-/issues/17993Address `Map Scalars` usability issues2021-03-23T18:11:22-04:00Utkarsh AyachitAddress `Map Scalars` usability issuesThe following needs to be done to make it less confusing when **Map Scalars** is checked/unchecked.
1. **Map Scalars** should only be un-checkable when direct coloring is possible i.e.
* for surface rendering, if the integral in range...The following needs to be done to make it less confusing when **Map Scalars** is checked/unchecked.
1. **Map Scalars** should only be un-checkable when direct coloring is possible i.e.
* for surface rendering, if the integral in range [0,255] or floating point in range [0.0, 1.0] and has number components between 1 and 4 (see `vtkScalarToColors` for interpretation for various components when using `VTK_COLOR_MODE_DIRECT_SCALARS`.
* for volume rendering, only 2 and 4 components would be supported with component interpretation matching that of surface rendering.
2. When **Map Scalars** is unchecked, the following should happen:
* the **color selection** combo-box in the toolbar and on the **Display** properties panel, should disable the component selection. Components cannot be selected when map-scalars is off.
* The **Color Map** editor should not show anything. There's no transfer function being set/used.
* The color legend should not be shown.https://gitlab.kitware.com/paraview/paraview/-/issues/17992Multiblock with normals are displayed incorrectly with vtkBlockColors with in...2021-03-23T18:12:10-04:00Mathieu Westphal (Kitware)Multiblock with normals are displayed incorrectly with vtkBlockColors with intel driverHere is how Multiblock with Normals are shown when showing the vtkBlockColors:
![mb](/uploads/9cd5f0bcf0b08ba9c7f9b9592ac54756/mb.png)
How to reproduce:
* Open ParaView
* Create a Sphere, Apply
* Group DataSet, Apply
* Show vtkBloc...Here is how Multiblock with Normals are shown when showing the vtkBlockColors:
![mb](/uploads/9cd5f0bcf0b08ba9c7f9b9592ac54756/mb.png)
How to reproduce:
* Open ParaView
* Create a Sphere, Apply
* Group DataSet, Apply
* Show vtkBlockColors -> incorrect rendering
* Add a PassArray, uncheck "Normals", Apply
* Show vtkBlockColors -> correct rendering
broken between ParaView 5.3.0 and ParaView 5.4.0,, exactly by vtk/vtk!2777.
Adding a test for this maybe a good idea.https://gitlab.kitware.com/paraview/paraview/-/issues/17989Select block should offer to select points2018-02-13T19:23:33-05:00W. Alan ScottSelect block should offer to select pointsSelect block should allow users to select either the block by cells or block by points.Select block should allow users to select either the block by cells or block by points.https://gitlab.kitware.com/paraview/paraview/-/issues/17984Plots not showing new arrays when array change.2018-02-13T19:18:33-05:00Utkarsh AyachitPlots not showing new arrays when array change.Steps:
1. open can.ex2, load only "EQPS" and "Global Element Ids". Apply
2. Create **Plot Data Over Time**. Apply. -- nothing will be shown as expected since there are no point array to plot.
3. Next change **Field Association** on the *...Steps:
1. open can.ex2, load only "EQPS" and "Global Element Ids". Apply
2. Create **Plot Data Over Time**. Apply. -- nothing will be shown as expected since there are no point array to plot.
3. Next change **Field Association** on the **Plot Data Over Time** filter to **Cells**, and hit Apply. Nothing shows up still. Nor is there anything on the Display panel that lets me choose any array to plot. That's bug. The only way to fix it is create a new view.https://gitlab.kitware.com/paraview/paraview/-/issues/17981Plot Selection Over Time fails for distributed datasets2020-09-23T15:50:15-04:00Utkarsh AyachitPlot Selection Over Time fails for distributed datasets1. Start ParaView, and connect to parallel pvserver (at least 2 ranks)
2. Open can.ex2, load only EQPS (just to keep things simple), Apply
3. D3, Apply, Color by "vtkProcessId" to see what's on each of the ranks.
4. Now select a few cell...1. Start ParaView, and connect to parallel pvserver (at least 2 ranks)
2. Open can.ex2, load only EQPS (just to keep things simple), Apply
3. D3, Apply, Color by "vtkProcessId" to see what's on each of the ranks.
4. Now select a few cells such that you're selecting cell on both ranks.
5. Plot Selection Over Time, Apply
The filter produces the following:
![image](/uploads/1f5b4c77b4fb43a50625232007e56ba6/image.png)
What we're seeing is that it created separate plots for data on each of the ranks. that's not what one would have expected.Yohann Bearzi (Kitware)Yohann Bearzi (Kitware)