From cory.quammen at kitware.com Wed Aug 2 16:01:12 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Wed, 2 Aug 2017 16:01:12 -0400 Subject: [Paraview-developers] ParaView 5.4.1-RC3 available for download Message-ID: On behalf of the ParaView development team, I am happy to announce that ParaView 5.4.1-RC3 is available for download. Binaries and source code are available from www.paraview.org/download The 5.4.1-RC3 bug fix release candidate features a couple improvements over 5.4.1-RC2: * A bug in the color mapping that in some cases incorrectly mapped some data values to the above-range color was fixed. * OSPRay can now be compiled in Cray programming environments and for the first time has support for Intel Skylake processors. Additional fixes since the release of ParaView 5.4.0 include: * The black render window observed on some systems when FXAA was enabled or an object's opacity was less than 1 has been fixed. * Field output for cells with non-straight edges has been corrected. * Derivatives for quadratic triangles are now computed correctly. * A bug that occurred when generating a Catalyst script with no images output has been fixed. * A problem with writing animations as AVIs was fixed. * Logic in the CGNS reader was updated to correctly read all solution nodes for each centering type when FlowSolutionPointers are missing. * Fixed crash in CGNS reader when moving forward in time. * Fixed problem with long-running animations not advancing past a certain number of time steps. * Resolved issue where changing one file path in the state file loader "Choose File Names" dialog also changed the file paths of other files with the same type. * Addressed problem with enumerated property dependencies in ParaViewWeb. * Added ability to change the font size in the ParaView user interface. * Fixed BoxLib3D file reader. * State files now properly support files with relative paths. * Removed inappropriate scalar bar scaling when saving large screen shots. Please report any problems you find with this release candidate on the mailing list or on the bug tracker at https://gitlab.kitware.com/paraview/paraview/issues Thanks! Cory -- Cory Quammen Staff R&D Engineer Kitware, Inc. From kriolog at gmail.com Wed Aug 2 16:08:20 2017 From: kriolog at gmail.com (Maxim Torgonskiy) Date: Wed, 2 Aug 2017 16:08:20 -0400 Subject: [Paraview-developers] bug in pqPipelineFilter Message-ID: Hello, It seems that there's a bug in pqPipelineFilter::getInputPortName(int). I have a vtk filter which has 3 inputs and when I try to access the inputs with pqPipelineFilter the function getInputPortName doesn't respect their order: .xml : .cxx : int Dummy::RequestData( vtkInformation *vtkNotUsed(info), vtkInformationVector **inputVector, vtkInformationVector *outputVector) { pqPipelineFilter* self_pqfilter = qobject_cast(pqActiveObjects::instance().activeSource()); Q_ASSERT(self_pqfilter); for(int i = 0; i < GetNumberOfInputPorts(); ++i) qDebug() << self_pqfilter->getInputPortName(i); return 1; } output : "One" "Two" "Zero" A minimal reproducible example is attached. Regards, Maxim -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Dummy.zip Type: application/zip Size: 2484 bytes Desc: not available URL: From utkarsh.ayachit at kitware.com Wed Aug 2 17:39:43 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Wed, 2 Aug 2017 17:39:43 -0400 Subject: [Paraview-developers] bug in pqPipelineFilter In-Reply-To: References: Message-ID: There's was never meant to be any order matching between the VTK port numbers and input ports provided by pqPipelineFilter. The latter indeed used a map internally using the property name as the key for the map (hence the change in order). If there's a convincing reason to preserve it, we can surely fix it, however. Utkarsh On Wed, Aug 2, 2017 at 4:08 PM, Maxim Torgonskiy wrote: > Hello, > > It seems that there's a bug in pqPipelineFilter::getInputPortName(int). I > have a vtk filter which has 3 inputs and when I try to access the inputs > with pqPipelineFilter the function getInputPortName doesn't respect their > order: > > .xml : > > > > > name="Zero" > port_index="0" > command="SetInputConnection"> > > > > > > > > > name="One" > port_index="1" > optional="1" > command="SetInputConnection"> > > > > > > > > > name="Two" > port_index="2" > optional="1" > command="SetInputConnection"> > > > > > > > > > > > > > .cxx : > int Dummy::RequestData( > vtkInformation *vtkNotUsed(info), > vtkInformationVector **inputVector, > vtkInformationVector *outputVector) > { > pqPipelineFilter* self_pqfilter = qobject_cast >(pqActiveObjects::instance().activeSource()); > Q_ASSERT(self_pqfilter); > > for(int i = 0; i < GetNumberOfInputPorts(); ++i) > qDebug() << self_pqfilter->getInputPortName(i); > > return 1; > } > > output : > "One" > "Two" > "Zero" > > > A minimal reproducible example is attached. > > Regards, > Maxim > > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Search the list archives at: http://markmail.org/search/?q= > Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From aron.helser at kitware.com Thu Aug 3 14:31:37 2017 From: aron.helser at kitware.com (Aron Helser) Date: Thu, 3 Aug 2017 14:31:37 -0400 Subject: [Paraview-developers] ParaView, Windows and QT 5.9.1 - no menubar? Message-ID: I recently starting compiling ParaView with Qt 5.9.1 on windows, and my menubars have disappeared. I can access them with keyboard shortcuts, but all my mouseclicks are offset vertically by about the size of the missing menubar. Has anyone else experienced this, or have suggestions? I'm not 100% sure it's QT 5.9.1, but it seems a likely suspect. [image: Inline image 1] Thanks, Aron -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: image.png Type: image/png Size: 153245 bytes Desc: not available URL: From cory.quammen at kitware.com Thu Aug 3 14:38:36 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Thu, 3 Aug 2017 14:38:36 -0400 Subject: [Paraview-developers] ParaView, Windows and QT 5.9.1 - no menubar? In-Reply-To: References: Message-ID: Ah ha, you can reproduce https://gitlab.kitware.com/paraview/paraview/issues/17499 ! We haven't made any headway as far as I know on that issue, mainly because we haven't been able to reproduce it. On Thu, Aug 3, 2017 at 2:31 PM, Aron Helser wrote: > I recently starting compiling ParaView with Qt 5.9.1 on windows, and my > menubars have disappeared. I can access them with keyboard shortcuts, but > all my mouseclicks are offset vertically by about the size of the missing > menubar. > > Has anyone else experienced this, or have suggestions? I'm not 100% sure > it's QT 5.9.1, but it seems a likely suspect. > > [image: Inline image 1] > Thanks, > Aron > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Search the list archives at: http://markmail.org/search/?q= > Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > > -- Cory Quammen Staff R&D Engineer Kitware, Inc. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: image.png Type: image/png Size: 153245 bytes Desc: not available URL: From ben.boeckel at kitware.com Thu Aug 3 14:49:14 2017 From: ben.boeckel at kitware.com (Ben Boeckel) Date: Thu, 3 Aug 2017 14:49:14 -0400 Subject: [Paraview-developers] ParaView, Windows and QT 5.9.1 - no menubar? In-Reply-To: References: Message-ID: <20170803184914.GA2313@megas.kitware.com> On Thu, Aug 03, 2017 at 14:31:37 -0400, Aron Helser wrote: > I recently starting compiling ParaView with Qt 5.9.1 on windows, and my > menubars have disappeared. I can access them with keyboard shortcuts, but > all my mouseclicks are offset vertically by about the size of the missing > menubar. > > Has anyone else experienced this, or have suggestions? I'm not 100% sure > it's QT 5.9.1, but it seems a likely suspect. Looks like Windows 10. I've noticed Firefox looking weird there with no title bar and the menus jammed along the top border. Maybe ParaView is doing that too, but then drawing the titlebar anyways? Is there some setting on the QMainWindow to make it behave properly? --Ben From aron.helser at kitware.com Thu Aug 3 15:40:05 2017 From: aron.helser at kitware.com (Aron Helser) Date: Thu, 3 Aug 2017 15:40:05 -0400 Subject: [Paraview-developers] ParaView, Windows and QT 5.9.1 - no menubar? In-Reply-To: <20170803184914.GA2313@megas.kitware.com> References: <20170803184914.GA2313@megas.kitware.com> Message-ID: Yes, I'm on Windows 10, compiling with VS2015. Firefox is mimicking the other browsers on Win10 - Chrome and Edge are similar. I was using QT 5.8.0 previously, and didn't have the problem. Hopefully this will help someone get a repro.... On Thu, Aug 3, 2017 at 2:49 PM, Ben Boeckel wrote: > On Thu, Aug 03, 2017 at 14:31:37 -0400, Aron Helser wrote: > > I recently starting compiling ParaView with Qt 5.9.1 on windows, and my > > menubars have disappeared. I can access them with keyboard shortcuts, but > > all my mouseclicks are offset vertically by about the size of the missing > > menubar. > > > > Has anyone else experienced this, or have suggestions? I'm not 100% sure > > it's QT 5.9.1, but it seems a likely suspect. > > Looks like Windows 10. I've noticed Firefox looking weird there with no > title bar and the menus jammed along the top border. Maybe ParaView is > doing that too, but then drawing the titlebar anyways? Is there some > setting on the QMainWindow to make it behave properly? > > --Ben > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cory.quammen at kitware.com Thu Aug 3 16:04:38 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Thu, 3 Aug 2017 16:04:38 -0400 Subject: [Paraview-developers] ParaView, Windows and QT 5.9.1 - no menubar? In-Reply-To: References: <20170803184914.GA2313@megas.kitware.com> Message-ID: Aron, Could you try downloading the 5.3 binary and see if you have the same issue? It uses Qt 5.8.0. I have Windows 10 on a low-resolution laptop and none of the 5.4.0 or 5.4.1 release candidates exhibit this behavior. Those builds use Qt 5.8.0. Ken Martin reports he has Windows 10 and has built ParaView against Qt 5.9.1, but the menu bar shows up just fine. Thanks, Cory On Thu, Aug 3, 2017 at 3:40 PM, Aron Helser wrote: > Yes, I'm on Windows 10, compiling with VS2015. Firefox is mimicking the > other browsers on Win10 - Chrome and Edge are similar. > I was using QT 5.8.0 previously, and didn't have the problem. > > Hopefully this will help someone get a repro.... > > On Thu, Aug 3, 2017 at 2:49 PM, Ben Boeckel wrote: >> >> On Thu, Aug 03, 2017 at 14:31:37 -0400, Aron Helser wrote: >> > I recently starting compiling ParaView with Qt 5.9.1 on windows, and my >> > menubars have disappeared. I can access them with keyboard shortcuts, >> > but >> > all my mouseclicks are offset vertically by about the size of the >> > missing >> > menubar. >> > >> > Has anyone else experienced this, or have suggestions? I'm not 100% sure >> > it's QT 5.9.1, but it seems a likely suspect. >> >> Looks like Windows 10. I've noticed Firefox looking weird there with no >> title bar and the menus jammed along the top border. Maybe ParaView is >> doing that too, but then drawing the titlebar anyways? Is there some >> setting on the QMainWindow to make it behave properly? >> >> --Ben > > -- Cory Quammen Staff R&D Engineer Kitware, Inc. From kriolog at gmail.com Fri Aug 4 09:52:14 2017 From: kriolog at gmail.com (Maxim Torgonskiy) Date: Fri, 4 Aug 2017 09:52:14 -0400 Subject: [Paraview-developers] bug in pqPipelineFilter In-Reply-To: References:

Message-ID: Sounds convincing, I'll switch to named access. Thank you for the clarification! Le 3 ao?t 2017 11:14, "Utkarsh Ayachit" a ?crit : While the mismatch may be confusing, ParaView deliberately opted for "named" access to the inputs since the proxy mechanism does provide ability to hide/ignore/change input ports. For example, proxy may not even have input properties for all input ports if some of them are optional. On Wed, Aug 2, 2017 at 6:19 PM, Maxim Torgonskiy wrote: > Ok, I see. I just supposed that the order of ports is the same. Before, in > my paraview plugin I used everywhere access by index (for both VTK and pqPipeline* > parts) and now I should mix it with access by input name which is less > consistent. > > Thanks, > Maxim > > > > 2017-08-02 17:39 GMT-04:00 Utkarsh Ayachit : > >> There's was never meant to be any order matching between the VTK port >> numbers and input ports provided by pqPipelineFilter. The latter indeed >> used a map internally using the property name as the key for the map >> (hence the change in order). >> >> If there's a convincing reason to preserve it, we can surely fix it, >> however. >> >> Utkarsh >> >> On Wed, Aug 2, 2017 at 4:08 PM, Maxim Torgonskiy >> wrote: >> >>> Hello, >>> >>> It seems that there's a bug in pqPipelineFilter::getInputPortName(int). >>> I have a vtk filter which has 3 inputs and when I try to access the inputs >>> with pqPipelineFilter the function getInputPortName doesn't respect their >>> order: >>> >>> .xml : >>> >>> >>> >>> >>> >> name="Zero" >>> port_index="0" >>> command="SetInputConnection"> >>> >>> >>> >>> >>> >>> >>> >>> >>> >> name="One" >>> port_index="1" >>> optional="1" >>> command="SetInputConnection"> >>> >>> >>> >>> >>> >>> >>> >>> >>> >> name="Two" >>> port_index="2" >>> optional="1" >>> command="SetInputConnection"> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> .cxx : >>> int Dummy::RequestData( >>> vtkInformation *vtkNotUsed(info), >>> vtkInformationVector **inputVector, >>> vtkInformationVector *outputVector) >>> { >>> pqPipelineFilter* self_pqfilter = qobject_cast>> >(pqActiveObjects::instance().activeSource()); >>> Q_ASSERT(self_pqfilter); >>> >>> for(int i = 0; i < GetNumberOfInputPorts(); ++i) >>> qDebug() << self_pqfilter->getInputPortName(i); >>> >>> return 1; >>> } >>> >>> output : >>> "One" >>> "Two" >>> "Zero" >>> >>> >>> A minimal reproducible example is attached. >>> >>> Regards, >>> Maxim >>> >>> >>> _______________________________________________ >>> Powered by www.kitware.com >>> >>> Visit other Kitware open-source projects at >>> http://www.kitware.com/opensource/opensource.html >>> >>> Search the list archives at: http://markmail.org/search/?q= >>> Paraview-developers >>> >>> Follow this link to subscribe/unsubscribe: >>> http://public.kitware.com/mailman/listinfo/paraview-developers >>> >>> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From amine.aboufirass at gmail.com Fri Aug 4 10:29:38 2017 From: amine.aboufirass at gmail.com (Amine Aboufirass) Date: Fri, 4 Aug 2017 16:29:38 +0200 Subject: [Paraview-developers] Control paraview interactively using IDLE? In-Reply-To: References:

Message-ID: Hi Cory, Thanks for your response. That is quite unfortunate. I think it would be very helpful for new, inexperienced users to "play around" with the interface in such a manner. Of course, I am one of these inexperienced users and so I have no idea of the development effort it would take to be able to do such a thing. On another note, I would like to ask about python versioning for the *existing* Paraview API. As you might know python is up to version 3.6 now. As far as I'm aware Paraview.simple works only on Python 2.7? Most libraries are now using at least 3.4 or so, especially libraries I would like to use Paraview with. Is an upgrade of the Paraview module underway? If not would it be possible for a lambda user to recompile them into later versions of python and use them? I had one guy from kitware tell me the following: No, recompilation is required to have the Python extensions using C to use > Python3. The easiest fix right now is to recompile ParaView yourself, sorry Of course I can try to recompile it but this would be a huge pain. So I don't want to embark on this adventure unless it is feasible within a reasonable time frame (which I doubt).... What's your take? Is it possible to use the API with Python 3? Regards, Amine On Sat, Jul 22, 2017 at 3:44 PM, Cory Quammen wrote: > Amine, > > As far as I am aware, I'm afraid that you will not be able to control > the paraview UI from an external Python. The ParaView GUI owns its > internal Python environment, and the Python environment (including the > paraview.simple module) does not have a direct reference to the > ParaView GUI side of things. It may appear that you are controlling > the GUI via Python in ParaView's Python console, but that happens > indirectly. Your Python script change objects in ParaView, but the UI > changes you see are the result of observers in the UI listening for > updates to ParaView's state and updating themselves based on the new > values. The Python script does not control the UI elements directly. > > The reason you see a separate render window is because running pure > Python scripts through pvpython or pvbatch, which are basically Python > executables with some additional path information, is explicitly > supported - the lone window appears when rendering commands are issued > in the script. What you have done with an external (and compatible) > Python is also supported, but it does not let you invoke the GUI > application, let alone control it. > > Hope that helps, > Cory > > On Sat, Jul 22, 2017 at 2:02 AM, Amine Aboufirass > wrote: > > I am trying to control Paraview interactively using IDLE. This would > involve > > sending commands from IDLE and seeing the changes occur in Paraview. I > would > > rather not use the in-Paraview python shell. > > > > So far, I have succeeded in importing the Paraview modules (simple, > > servermanager?etc) from IDLE. However the commands sent do not reflect in > > Paraview. For instance: > > > >>>> from paraview.simple import * > >>>> cone = Cone() > >>>> Show() > >>>> Render() > > > > does indeed create a cone. However the cone is output to a new, > independent > > OpenGL window, and not the Paraview GUI. > > > > Is it possible to control Paraview interactively using IDLE? If so how to > > accomplish this? Thanks > > > > > > _______________________________________________ > > Powered by www.kitware.com > > > > Visit other Kitware open-source projects at > > http://www.kitware.com/opensource/opensource.html > > > > Search the list archives at: > > http://markmail.org/search/?q=Paraview-developers > > > > Follow this link to subscribe/unsubscribe: > > http://public.kitware.com/mailman/listinfo/paraview-developers > > > > > > -- > Cory Quammen > Staff R&D Engineer > Kitware, Inc. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cory.quammen at kitware.com Fri Aug 4 10:38:20 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Fri, 4 Aug 2017 10:38:20 -0400 Subject: [Paraview-developers] Control paraview interactively using IDLE? In-Reply-To: References:

Message-ID: Amine, No, you won't be able to use the Python API from within a binary you have downloaded from paraview.org with Python 3. You would indeed need to compile ParaView with Python 3. Doing so is officially supported (we have regular test builds against Python 3), we just don't ship the binaries with Python 3 for now. I'm not sure when we will ship binaries with Python 3, perhaps in the next couple releases. Best regards, Cory On Fri, Aug 4, 2017 at 10:29 AM, Amine Aboufirass wrote: > Hi Cory, > > Thanks for your response. That is quite unfortunate. I think it would be > very helpful for new, inexperienced users to "play around" with the > interface in such a manner. Of course, I am one of these inexperienced users > and so I have no idea of the development effort it would take to be able to > do such a thing. > > On another note, I would like to ask about python versioning for the > existing Paraview API. As you might know python is up to version 3.6 now. As > far as I'm aware Paraview.simple works only on Python 2.7? Most libraries > are now using at least 3.4 or so, especially libraries I would like to use > Paraview with. > > Is an upgrade of the Paraview module underway? If not would it be possible > for a lambda user to recompile them into later versions of python and use > them? > > I had one guy from kitware tell me the following: > >> No, recompilation is required to have the Python extensions using C to use >> Python3. The easiest fix right now is to recompile ParaView yourself, sorry > > > Of course I can try to recompile it but this would be a huge pain. So I > don't want to embark on this adventure unless it is feasible within a > reasonable time frame (which I doubt).... > > What's your take? Is it possible to use the API with Python 3? > > Regards, > > Amine > > > On Sat, Jul 22, 2017 at 3:44 PM, Cory Quammen > wrote: >> >> Amine, >> >> As far as I am aware, I'm afraid that you will not be able to control >> the paraview UI from an external Python. The ParaView GUI owns its >> internal Python environment, and the Python environment (including the >> paraview.simple module) does not have a direct reference to the >> ParaView GUI side of things. It may appear that you are controlling >> the GUI via Python in ParaView's Python console, but that happens >> indirectly. Your Python script change objects in ParaView, but the UI >> changes you see are the result of observers in the UI listening for >> updates to ParaView's state and updating themselves based on the new >> values. The Python script does not control the UI elements directly. >> >> The reason you see a separate render window is because running pure >> Python scripts through pvpython or pvbatch, which are basically Python >> executables with some additional path information, is explicitly >> supported - the lone window appears when rendering commands are issued >> in the script. What you have done with an external (and compatible) >> Python is also supported, but it does not let you invoke the GUI >> application, let alone control it. >> >> Hope that helps, >> Cory >> >> On Sat, Jul 22, 2017 at 2:02 AM, Amine Aboufirass >> wrote: >> > I am trying to control Paraview interactively using IDLE. This would >> > involve >> > sending commands from IDLE and seeing the changes occur in Paraview. I >> > would >> > rather not use the in-Paraview python shell. >> > >> > So far, I have succeeded in importing the Paraview modules (simple, >> > servermanager?etc) from IDLE. However the commands sent do not reflect >> > in >> > Paraview. For instance: >> > >> >>>> from paraview.simple import * >> >>>> cone = Cone() >> >>>> Show() >> >>>> Render() >> > >> > does indeed create a cone. However the cone is output to a new, >> > independent >> > OpenGL window, and not the Paraview GUI. >> > >> > Is it possible to control Paraview interactively using IDLE? If so how >> > to >> > accomplish this? Thanks >> > >> > >> > _______________________________________________ >> > Powered by www.kitware.com >> > >> > Visit other Kitware open-source projects at >> > http://www.kitware.com/opensource/opensource.html >> > >> > Search the list archives at: >> > http://markmail.org/search/?q=Paraview-developers >> > >> > Follow this link to subscribe/unsubscribe: >> > http://public.kitware.com/mailman/listinfo/paraview-developers >> > >> >> >> >> -- >> Cory Quammen >> Staff R&D Engineer >> Kitware, Inc. > > -- Cory Quammen Staff R&D Engineer Kitware, Inc. From wascott at sandia.gov Fri Aug 4 13:56:39 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Fri, 4 Aug 2017 17:56:39 +0000 Subject: [Paraview-developers] 64 bit element IDs Message-ID: Does ParaView support handling of 64 bit node and element IDs? If not, we need to add it. Sandia is seeing datasets that are failing with 32 bit IDs. If necessary, I have a user that can create datasets that show this problem. Thanks, Alan -------------------------------------------------------- W. Alan Scott ParaView Support Manager SAIC Sandia National Laboratories, MS 0822 Org 9326 - Building 880 A1-K (505) 284-0932 FAX (505) 284-5619 The most exciting phrase to hear in science is not "Eureka!" but "That's funny..." -- Isaac Asimov --------------------------------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From cory.quammen at kitware.com Fri Aug 4 14:22:12 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Fri, 4 Aug 2017 14:22:12 -0400 Subject: [Paraview-developers] 64 bit element IDs In-Reply-To: References: Message-ID: On Fri, Aug 4, 2017 at 1:56 PM, Scott, W Alan wrote: > Does ParaView support handling of 64 bit node and element IDs? Yes, it can. If you have data with 64-bit IDs that doesn't work, please let us know. Thanks, Cory > If not, we > need to add it. Sandia is seeing datasets that are failing with 32 bit IDs. > If necessary, I have a user that can create datasets that show this problem. > > > > Thanks, > > > > Alan > > > > > > -------------------------------------------------------- > > W. Alan Scott > > ParaView Support Manager > > > > SAIC > > Sandia National Laboratories, MS 0822 > > Org 9326 - Building 880 A1-K > > (505) 284-0932 FAX (505) 284-5619 > > > > The most exciting phrase to hear in science > > is not "Eureka!" but "That's funny..." -- Isaac Asimov > > --------------------------------------------------------- > > > > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at > http://www.kitware.com/opensource/opensource.html > > Search the list archives at: > http://markmail.org/search/?q=Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > -- Cory Quammen Staff R&D Engineer Kitware, Inc. From cmpute at foxmail.com Fri Aug 4 21:31:52 2017 From: cmpute at foxmail.com (=?gb18030?B?1tPUqvbO?=) Date: Sat, 5 Aug 2017 09:31:52 +0800 Subject: [Paraview-developers] About the implementation of filter monitoring and selection extraction Message-ID: Hi, I have some questions of the filter structure in ParaView. I'm developing a client based on ParaView, and I want to wrap my app so that the filters are not directly exposed. Instead, I want to expose certain filters as customize objects. So there is a question, what should I do to monitor the changes of filters? In addition, I wonder how ExtractSelection filter is connected with selecition node in pqRenderView because I'm going to implement customized selection tool in my app. I found that there is a property call "SelectionInput" in ParaView xml configuration, but I didn't found it elsewhere in the code that seems to use the property and connect selection to the filter. Can anyone help me with how the manager s listen to the change of filters and connect selection to extractor? Thanks, Jacob Zhong -------------- next part -------------- An HTML attachment was scrubbed... URL: From bobpepin at gmail.com Sun Aug 6 17:04:59 2017 From: bobpepin at gmail.com (Bob Pepin) Date: Sun, 06 Aug 2017 21:04:59 +0000 Subject: [Paraview-developers] Announce: vtkPythonClassAlgorithm Message-ID: Hi, I developed a ParaView plugin that is similar to Programmable Filter/Source, except that the Python code is in a Python class in an external file. The source code for a first (rough) version can be found on github at https://github.com/bobpepin/vtkPythonClassAlgorithm I would be thankful for feedback or suggestions (as well as contributions of course). Best, Bob -------------- next part -------------- An HTML attachment was scrubbed... URL: From bobpepin at gmail.com Sun Aug 6 17:20:28 2017 From: bobpepin at gmail.com (Bob Pepin) Date: Sun, 06 Aug 2017 21:20:28 +0000 Subject: [Paraview-developers] Control paraview interactively using IDLE? Message-ID: (sorry for out-of-thread message, I only joined the mailing list today and saw this in the archives) Hi Amine, I have had some success controlling paraview from Python using multi-client mode. Basically you need to start pvserver using the --multi-clients option, then connect to the same server both with a GUI and a Python client. You can now create objects from the Python client and have them appear in the GUI, and there is even limited support for playing around with Layouts and such. However, I have to say that it was quite easy to crash paraview when playing around with this. My original motivation for doing this was to see if I could replace Matlab with Python plus some plotting software that can zoom, pick points and do some 3d (interactivity in the common Python plotting packages is surprisingly limited). ParaView looks promising but I haven't pursued the interactive part much since my initial experiments with multi-client mode, for me it worked well enough just to write data out to hdf5 files and create little xdmf3 files to go with it. Hope this helps, Bob -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Sun Aug 6 22:36:38 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Sun, 6 Aug 2017 22:36:38 -0400 Subject: [Paraview-developers] Announce: vtkPythonClassAlgorithm In-Reply-To: References: Message-ID: Bob, That's pretty cool! I'll take a closer look at it next week. This seems to be in similar vein to a WIP-branch I created a little while ago but never got around to finishing it. If you're interested, here's the commit: https://gitlab.kitware.com/utkarsh.ayachit/paraview/commit/bfea05b0d2114f51a581f63c368cdce603932619 It relies on `vtkPythonAlgorithm` ( https://blog.kitware.com/vtkpythonalgorithm-is-great/) to handle passing algorithm passes to a Python class. The XML looks like Where the "class" is the Python class name indeed with extra being the module to import. Utkarsh On Sun, Aug 6, 2017 at 5:04 PM, Bob Pepin wrote: > Hi, > I developed a ParaView plugin that is similar to Programmable > Filter/Source, except that the Python code is in a Python class in an > external file. > > The source code for a first (rough) version can be found on github at > https://github.com/bobpepin/vtkPythonClassAlgorithm > > I would be thankful for feedback or suggestions (as well as contributions > of course). > > Best, > Bob > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Search the list archives at: http://markmail.org/search/?q= > Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jonathan.borduas at caboma.com Mon Aug 7 11:40:53 2017 From: jonathan.borduas at caboma.com (Jonathan Borduas) Date: Mon, 7 Aug 2017 15:40:53 +0000 Subject: [Paraview-developers] Rendering/Screenshot with 32 bit depth colors (32 bpp) Message-ID: Dear Paraviewers, Is there a setting available to configure the bit depth (Bit Per Pixel - bpp) in ParaView ? My goal is to render and export images that require more than the 24 bits depth. Ideally 32 bits. I have an intel hd graphics 4000 on Windows 10 which support Intel's Deep Color 32 bit. P.S. volumic rendering isn't working on Intel hd graphics 4000, maybe because it uses 32 bits for rendering and Paraview doesn't ? Best regards, Jonathan Borduas -------------- next part -------------- An HTML attachment was scrubbed... URL: From bobpepin at gmail.com Mon Aug 7 14:02:45 2017 From: bobpepin at gmail.com (Bob Pepin) Date: Mon, 07 Aug 2017 18:02:45 +0000 Subject: [Paraview-developers] Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: Hi Utkarsh, thank you for sharing your code, I think the fact that you started work on something similar shows that there is indeed a need for this kind of feature. Would there be an interest in integrating this with ParaView once it has reached a certain maturity? Your approach seems to be complementary to what I have implemented, it seems to me you could easily replace vtkPythonAlgo with vtkPythonClassAlgo. I think that having a single string for module.class is indeed the better solution, I went with separate parameters mostly because it was faster to implement. In particular, Py_ImportModule("foo.bar.baz") returns the module object for foo, and we will have to traverse to baz manually if we want to create an instance from a class in baz. I also think that a vanilla vtkPythonAlgorithm can not meet all the needs of this feature. In my case I need for example to override processRequest on the C++ level in order to reload the module on a new request (and ideally only when the source code has changed, and only optionally). The main issue going forward I can see right now involves passing parameters to Python. Right now I have a C++ method SetStringProperty(name, value) and use a with 2 elements. Now suppose I want to have a FileName property so that SetStringProperty("filename", value) is called when the FileName property is set to value. Do you know how to achieve this with the existing mechanisms, for example is it possible to have a StringVectorProperty with 2 elements where the first one is hidden and set to a default value? If not, do you think it would be possible to pass the value of a StringVectorProperty through sprintf with a custom format string before calling the command? For example, so that would result in the method call SetProperty("filename=foo.bar") when the FileName property is set to "foo.bar". Best, Bob On Mon, Aug 7, 2017 at 4:36 AM Utkarsh Ayachit wrote: > Bob, > > That's pretty cool! I'll take a closer look at it next week. This seems to > be in similar vein to a WIP-branch I created a little while ago but never > got around to finishing it. If you're interested, here's the commit: > > > https://gitlab.kitware.com/utkarsh.ayachit/paraview/commit/bfea05b0d2114f51a581f63c368cdce603932619 > > It relies on `vtkPythonAlgorithm` ( > https://blog.kitware.com/vtkpythonalgorithm-is-great/) to handle passing > algorithm passes to a Python class. The XML looks like > > name="PythonAlgorithm" > si_class="vtkSIPythonSourceProxy"> > > > Where the "class" is the Python class name indeed with extra being the > module to import. > > Utkarsh > > > On Sun, Aug 6, 2017 at 5:04 PM, Bob Pepin wrote: > >> Hi, >> I developed a ParaView plugin that is similar to Programmable >> Filter/Source, except that the Python code is in a Python class in an >> external file. >> >> The source code for a first (rough) version can be found on github at >> https://github.com/bobpepin/vtkPythonClassAlgorithm >> >> I would be thankful for feedback or suggestions (as well as contributions >> of course). >> >> Best, >> Bob >> >> _______________________________________________ >> Powered by www.kitware.com >> >> Visit other Kitware open-source projects at >> http://www.kitware.com/opensource/opensource.html >> >> Search the list archives at: >> http://markmail.org/search/?q=Paraview-developers >> >> Follow this link to subscribe/unsubscribe: >> http://public.kitware.com/mailman/listinfo/paraview-developers >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cory.quammen at kitware.com Fri Aug 11 16:52:02 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Fri, 11 Aug 2017 16:52:02 -0400 Subject: [Paraview-developers] ParaView 5.4.1-RC4 available for download Message-ID: On behalf of the ParaView development team, I am pleased to announce that ParaView 5.4.1-RC4 is available for download. Binaries and source code are available from www.paraview.org/download The 5.4.1-RC4 bug fix release candidate sports a couple of improvements since 5.4.1-RC3: * Added lower limit to how small the GUI font size can be. * Addressed issue affecting layouts when loading a state file from the Python console. Please report any problems you find with this release candidate on the mailing list or on the bug tracker at https://gitlab.kitware.com/paraview/paraview/issues Note: RC4 is expected to be the last release candidate in the 5.4.1 release cycle. Thanks! -- Cory Quammen Staff R&D Engineer Kitware, Inc. From wascott at sandia.gov Mon Aug 14 18:39:11 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Mon, 14 Aug 2017 22:39:11 +0000 Subject: [Paraview-developers] [EXTERNAL] Re: Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: Does this mean the Python could be developed in an IDE? Alan From: Paraview-developers [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Utkarsh Ayachit Sent: Sunday, August 6, 2017 8:37 PM To: Bob Pepin Cc: ParaView Developers Subject: [EXTERNAL] Re: [Paraview-developers] Announce: vtkPythonClassAlgorithm Bob, That's pretty cool! I'll take a closer look at it next week. This seems to be in similar vein to a WIP-branch I created a little while ago but never got around to finishing it. If you're interested, here's the commit: https://gitlab.kitware.com/utkarsh.ayachit/paraview/commit/bfea05b0d2114f51a581f63c368cdce603932619 It relies on `vtkPythonAlgorithm` (https://blog.kitware.com/vtkpythonalgorithm-is-great/) to handle passing algorithm passes to a Python class. The XML looks like Where the "class" is the Python class name indeed with extra being the module to import. Utkarsh On Sun, Aug 6, 2017 at 5:04 PM, Bob Pepin > wrote: Hi, I developed a ParaView plugin that is similar to Programmable Filter/Source, except that the Python code is in a Python class in an external file. The source code for a first (rough) version can be found on github at https://github.com/bobpepin/vtkPythonClassAlgorithm I would be thankful for feedback or suggestions (as well as contributions of course). Best, Bob _______________________________________________ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Search the list archives at: http://markmail.org/search/?q=Paraview-developers Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview-developers -------------- next part -------------- An HTML attachment was scrubbed... URL: From hahnse at ornl.gov Mon Aug 14 21:26:20 2017 From: hahnse at ornl.gov (Hahn, Steven E.) Date: Tue, 15 Aug 2017 01:26:20 +0000 Subject: [Paraview-developers] Connecting Minimum/Maximum from threshold filter and color map range Message-ID: <1502760379724.92667@ornl.gov> I want to add an option to adjust the color map range so that it stays between the minimum and maximum specified by the threshold filter. How can I connect these two items? Best, Steven -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Mon Aug 14 22:36:20 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Mon, 14 Aug 2017 22:36:20 -0400 Subject: [Paraview-developers] Connecting Minimum/Maximum from threshold filter and color map range In-Reply-To: <1502760379724.92667@ornl.gov> References: <1502760379724.92667@ornl.gov> Message-ID: Steven, How about monitoring the `vtkCommand::ModifiedEvent` on the `ThresholdBetween` property on the filter and then calling `vtkSMTransferFunctionProxy::RescaleTransferFunction` in the callback? Utkarsh On Mon, Aug 14, 2017 at 9:26 PM, Hahn, Steven E. wrote: > I want to add an option to adjust the color map range so that it stays > between the minimum and maximum specified by the threshold filter. How can > I connect these two items? > > > Best, > > Steven > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Search the list archives at: http://markmail.org/search/?q= > Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Mon Aug 14 22:42:00 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Mon, 14 Aug 2017 22:42:00 -0400 Subject: [Paraview-developers] Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: Bob, > Would there be an interest in integrating this with ParaView once it has reached a certain maturity? Surely. There's value to such a capability, no doubt. > I also think that a vanilla vtkPythonAlgorithm can not meet all the needs of this feature. In my case I need for example to override processRequest on the C++ level in order to reload the module on a new request (and ideally only when the source code has changed, and only optionally). This is a use-case that I'm skeptical we can support robustly in ParaView esp. in all modes it operates (client-server etc). I am not convinced it's that useful besides debugging, however. We can sure add a "reload" button to manually reimport the module, but that's not exactly what you want. But I'll stay posted. > The main issue going forward I can see right now involves passing parameters to Python. Right now I have a C++ method SetStringProperty(name, value) and use a with 2 elements. Now suppose I want to have a FileName property so that SetStringProperty("filename", value) is called when the FileName property is set to value. Do you know how to achieve this with the existing mechanisms, for example is it possible to have a StringVectorProperty with 2 elements where the first one is hidden and set to a default value? Are you aware of this post: https://blog.kitware.com/easy-customization-of-the-paraview-python-programmable-filter-property-panel/ Utkarsh From utkarsh.ayachit at kitware.com Mon Aug 14 22:43:09 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Mon, 14 Aug 2017 22:43:09 -0400 Subject: [Paraview-developers] [EXTERNAL] Re: Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: > Does this mean the Python could be developed in an IDE? Sure, this will support that. From bobpepin at gmail.com Wed Aug 16 07:37:22 2017 From: bobpepin at gmail.com (Bob Pepin) Date: Wed, 16 Aug 2017 11:37:22 +0000 Subject: [Paraview-developers] [EXTERNAL] Re: Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: Hi Alan, since this is a plugin, it means in fact that Python extensions can be developed in an IDE right now, even for older ParaView versions (assuming you have an environment in which to compile ParaView modules). If you want to give it a try, I would be very interested to hear if you got it to work on your platform or whether you had any trouble with the plugin or the documentation. Best, Bob On Tue, Aug 15, 2017 at 4:43 AM Utkarsh Ayachit wrote: > > Does this mean the Python could be developed in an IDE? > > Sure, this will support that. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From bobpepin at gmail.com Wed Aug 16 07:50:31 2017 From: bobpepin at gmail.com (Bob Pepin) Date: Wed, 16 Aug 2017 11:50:31 +0000 Subject: [Paraview-developers] Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: > > Hi, > > > I also think that a vanilla vtkPythonAlgorithm can not meet all the > needs of this feature. In my case I need for example to override > processRequest on the C++ level in order to reload the module on a new > request (and ideally only when the source code has changed, and only > optionally). > > This is a use-case that I'm skeptical we can support robustly in > ParaView esp. in all modes it operates (client-server etc). I am not > convinced it's that useful besides debugging, however. We can sure add > a "reload" button to manually reimport the module, but that's not > exactly what you want. But I'll stay posted. > > do you have a specific situation in mind regarding the reloading? Since the module uses Python's native functionality for reloading modules, I would assume that if we can load it, we can reload it. I agree however that checking whether the source has changed is a bit more delicate. Besides, I think that a Reload button would be perfect, especially if it did an Apply at the same time. I was considering adding a checkbox that ends up calling Modified() on the vtk Algorithm object upon checking/unchecking, do you know of a better solution using only existing ParaView features? > > The main issue going forward I can see right now involves passing > parameters to Python. Right now I have a C++ method SetStringProperty(name, > value) and use a with 2 elements. Now suppose I want > to have a FileName property so that SetStringProperty("filename", value) is > called when the FileName property is set to value. Do you know how to > achieve this with the existing mechanisms, for example is it possible to > have a StringVectorProperty with 2 elements where the first one is hidden > and set to a default value? > > Are you aware of this post: > > https://blog.kitware.com/easy-customization-of-the-paraview-python-programmable-filter-property-panel/ Thanks, This was exactly what I was looking for. Just out of curiosity, how does the "magic" work here? I.e. how does PV decide whether to call SetProperty(value) or SetProperty(name, value)? I also added a HDF5 Reader plugin to the github repository as an example. Best, Bob -------------- next part -------------- An HTML attachment was scrubbed... URL: From wascott at sandia.gov Thu Aug 17 13:38:04 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Thu, 17 Aug 2017 17:38:04 +0000 Subject: [Paraview-developers] Conference (was Is anyone using the Extra Headlight in ParaView?) Message-ID: Tim, What conference were you attending? How many students attended? Any good positive or negative feedback? Did you use your own training notes, or one of the ones in the paraview.org/ training page? Ideas are always welcome! Alan From: Timothy Dwight Dunn [mailto:timothy.dunn at colorado.edu] Sent: Thursday, August 17, 2017 11:34 AM To: Scott, W Alan Cc: DeMarle, David E. (External Contacts) ; paraview at paraview.org Subject: Re: [Paraview] [EXTERNAL] Is anyone using the Extra Headlight in ParaView? Actually I was just talking about this in a Paraview tutorial I gave this morning at a conference. I would prefer to kill the hard coded light allowing the ability to add in general lights that the user can design and setup. In addition to save a conf out for the setup to be used as a default. I normally set things up in Paraview export what I need pop them in a CG app so that I can light them with so much more control but this is often difficult at best to impossible as the norm due to complexity of the scene. If we could get more and better light control, especially with Osprey then I'd be a very happy camper. Just to make really difficult make sure the lights interact with objects/layers as desired by the user, including volumeterics. This is a common thing to do in CG apps. Light objects A,B,C with light rig1, object D,E with light rig2 and object F with light rig3 .... The easiest way to do this is to have a panel for each light, or for paraview the gui controls with a check box list for each available layer which can activate/deactivate each light per layer. OK this makes sense in my head but I don't think I'm explaining it well do to conference exhaustion. If desired I can do a mock up gui layout image of what I'm trying to describe. Tim Dunn Scientific\Data Visualization & Visual Analytics for; CIRES - Earth Lab-Analytics Hub CU Research Computing (CURC) Center for Research Data and Digital Scholarship (CRDDS) tdunn at colorado.edu On Thu, Aug 17, 2017 at 11:16 AM, Scott, W Alan > wrote: No objection here. Only thought I have is that the lights may be setup by default slightly differently for 2d and 3d. We may want a ?headlight? for 2d. Alan From: ParaView [mailto:paraview-bounces at paraview.org] On Behalf Of David E DeMarle Sent: Thursday, August 17, 2017 7:35 AM To: paraview at paraview.org Subject: [EXTERNAL] [Paraview] Is anyone using the Extra Headlight in ParaView? By default, Paraview illuminates the scene with a vtkLightKit. You can however add an extra headlight. I am about to add a more general capability which lets you add any number and any kind of light, including the headlight. When I do that, the semi-hardcoded extra headlight becomes redundant. I would like to remove it to streamline both the code and the UI. Are there any objections to me doing that? David E DeMarle Kitware, Inc. Principal Engineer 21 Corporate Drive Clifton Park, NY 12065-8662 Phone: 518-881-4909 _______________________________________________ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView Search the list archives at: http://markmail.org/search/?q=ParaView Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview -------------- next part -------------- An HTML attachment was scrubbed... URL: From wascott at sandia.gov Fri Aug 18 16:33:26 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Fri, 18 Aug 2017 20:33:26 +0000 Subject: [Paraview-developers] [EXTERNAL] Re: Announce: vtkPythonClassAlgorithm In-Reply-To: References:

Message-ID: <70ebe2a9a540443f9388beca16f48c45@ES01AMSNLNT.srn.sandia.gov> Ping From: Paraview-developers [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Scott, W Alan Sent: Monday, August 14, 2017 4:39 PM To: Ayachit, Utkarsh (External Contacts) ; Bob Pepin Cc: ParaView Developers Subject: Re: [Paraview-developers] [EXTERNAL] Re: Announce: vtkPythonClassAlgorithm Does this mean the Python could be developed in an IDE? Alan From: Paraview-developers [mailto:paraview-developers-bounces at paraview.org] On Behalf Of Utkarsh Ayachit Sent: Sunday, August 6, 2017 8:37 PM To: Bob Pepin > Cc: ParaView Developers > Subject: [EXTERNAL] Re: [Paraview-developers] Announce: vtkPythonClassAlgorithm Bob, That's pretty cool! I'll take a closer look at it next week. This seems to be in similar vein to a WIP-branch I created a little while ago but never got around to finishing it. If you're interested, here's the commit: https://gitlab.kitware.com/utkarsh.ayachit/paraview/commit/bfea05b0d2114f51a581f63c368cdce603932619 It relies on `vtkPythonAlgorithm` (https://blog.kitware.com/vtkpythonalgorithm-is-great/) to handle passing algorithm passes to a Python class. The XML looks like Where the "class" is the Python class name indeed with extra being the module to import. Utkarsh On Sun, Aug 6, 2017 at 5:04 PM, Bob Pepin > wrote: Hi, I developed a ParaView plugin that is similar to Programmable Filter/Source, except that the Python code is in a Python class in an external file. The source code for a first (rough) version can be found on github at https://github.com/bobpepin/vtkPythonClassAlgorithm I would be thankful for feedback or suggestions (as well as contributions of course). Best, Bob _______________________________________________ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Search the list archives at: http://markmail.org/search/?q=Paraview-developers Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview-developers -------------- next part -------------- An HTML attachment was scrubbed... URL: From wascott at sandia.gov Fri Aug 18 18:38:42 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Fri, 18 Aug 2017 22:38:42 +0000 Subject: [Paraview-developers] Weird quirk in the Exodus reader in Python Message-ID: >From a user. Is this a bug or feature? Should it be written up? Note that when User talks about updates, pre-update was TOSS II, post update is TOSS III. (TOSS is the OS for our clusters). Hi Alan, I figured you might want to know what the fix to my issues is. When loading a partitioned exodus (.e) file (ex. Simulation.e.5.0, Simulation.e.5.1, Simulation.e.5.2 etc. etc.) the array that you load via ExodusIIReader must be sorted. This was not a requirement in previous Paraview versions, hence why my python script stopped working in newer versions. For example, if you have a partitioned exodus file which is in an unsorted list as follows: file_list = ['simulation.e.5.1', 'simulation.e.5.4', 'simulation.e.5.2', 'simulation.e.5.0', 'simulation.e.5.3'] Then the following command will work as intended in Paraview 5.1.2 (on pre-update systems like Chama): exo = ExodusIIReader(FileName = file_list) If you try the above command on 5.4.0 or 5.1.2 on a post-updated system (like Skybridge), it doesn't work (no blocks show) and you get the following error message when trying to use the ColorBy command: Warning: In /projects/viz/paraview/src/ParaView/5.4.0-sb/ParaView5/ParaViewCore/ServerManager/Rendering/vtkSMPVRepresentationProxy.cxx, line 276 vtkSMPVRepresentationProxy (0x4d8f0f0): Could not determine array range. If you instead use the following command: exo = ExodusIIReader(FileName = sorted(file_list)) There are no issues. Perhaps you're already aware of this quirk, but I thought I'd let you know just in case. -------------------------------------------------------- W. Alan Scott ParaView Support Manager SAIC Sandia National Laboratories, MS 0807 Org 9326 - Building 880 A1-K (505) 284-0932 FAX (505) 284-5619 The most exciting phrase to hear in science is not "Eureka!" but "That's funny..." -- Isaac Asimov --------------------------------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Mon Aug 21 10:34:35 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Mon, 21 Aug 2017 10:34:35 -0400 Subject: [Paraview-developers] Weird quirk in the Exodus reader in Python In-Reply-To: References: Message-ID: Alan, Can you share a dataset to reproduce this, it'll be easier to figure the issue out that way. If I was to make a guess, I suspect that the reader reads array info etc. from the first file, if the first file doesn't have all the info, it may not have things determined as expected -- but that's just a guess. Utkarsh On Fri, Aug 18, 2017 at 6:38 PM, Scott, W Alan wrote: > From a user. Is this a bug or feature? Should it be written up? Note that > when User talks about updates, pre-update was TOSS II, post update is TOSS > III. (TOSS is the OS for our clusters). > > > > > > Hi Alan, > > > > I figured you might want to know what the fix to my issues is. > > > > When loading a partitioned exodus (.e) file (ex. Simulation.e.5.0, > Simulation.e.5.1, Simulation.e.5.2 etc. etc.) the array that you load via > ExodusIIReader must be sorted. This was not a requirement in previous > Paraview versions, hence why my python script stopped working in newer > versions. > > > > For example, if you have a partitioned exodus file which is in an unsorted > list as follows: > > file_list = ['simulation.e.5.1', 'simulation.e.5.4', 'simulation.e.5.2', > 'simulation.e.5.0', 'simulation.e.5.3'] > > > > Then the following command will work as intended in Paraview 5.1.2 (on > pre-update systems like Chama): > > exo = ExodusIIReader(FileName = file_list) > > > > If you try the above command on 5.4.0 or 5.1.2 on a post-updated system > (like Skybridge), it doesn't work (no blocks show) and you get the following > error message when trying to use the ColorBy command: > > > > Warning: In > /projects/viz/paraview/src/ParaView/5.4.0-sb/ParaView5/ParaViewCore/ServerManager/Rendering/vtkSMPVRepresentationProxy.cxx, > line 276 vtkSMPVRepresentationProxy (0x4d8f0f0): Could not determine array > range. > > > > > > If you instead use the following command: > > exo = ExodusIIReader(FileName = sorted(file_list)) > > > > There are no issues. Perhaps you're already aware of this quirk, but I > thought I'd let you know just in case. > > > > > > > > -------------------------------------------------------- > > W. Alan Scott > > ParaView Support Manager > > > > SAIC > > Sandia National Laboratories, MS 0807 > > Org 9326 - Building 880 A1-K > > (505) 284-0932 FAX (505) 284-5619 > > > > The most exciting phrase to hear in science > > is not "Eureka!" but "That's funny..." -- Isaac Asimov > > --------------------------------------------------------- > > > > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at > http://www.kitware.com/opensource/opensource.html > > Search the list archives at: > http://markmail.org/search/?q=Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > From cory.quammen at kitware.com Mon Aug 21 11:04:16 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Mon, 21 Aug 2017 11:04:16 -0400 Subject: [Paraview-developers] ParaView 5.4.1 is now available for download Message-ID: The ParaView development team is pleased to announce that ParaView 5.4.1 is available for download. Binaries and source code are available from www.paraview.org/download Note that the ParaView Download page has been updated with an improved layout that makes downloading ParaView easier. The ParaView 5.4.1 bug fix release addresses a number of problems in 5.4.0, including: * The black render window observed on some systems when FXAA was enabled or an object's opacity was less than 1 has been fixed. * Field output for cells with non-straight edges has been corrected. * Derivatives for quadratic triangles are now computed correctly. * A bug that occurred when generating a Catalyst script with no images output has been fixed. * A problem with writing animations as AVIs was fixed. * Logic in the CGNS reader was updated to correctly read all solution nodes for each centering type when FlowSolutionPointers are missing. * Fixed crash in CGNS reader when moving forward in time. * Fixed problem with long-running animations not advancing past a certain number of time steps. * Resolved issue where changing one file path in the state file loader "Choose File Names" dialog also changed the file paths of other files with the same type. * Addressed problem with enumerated property dependencies in ParaViewWeb. * Added ability to change the font size in the ParaView user interface. * Fixed BoxLib3D file reader. * State files now properly support files with relative paths. * Removed inappropriate scalar bar scaling when saving large screen shots. * A bug in the color mapping that in some cases incorrectly mapped some data values to the above-range color was fixed. * OSPRay can now be compiled in Cray programming environments and for the first time has support for Intel Skylake processors. * Added lower limit to how small the GUI font size can be. * Addressed issue affecting layouts when loading a state file from the Python console. Thanks! -- Cory Quammen Staff R&D Engineer Kitware, Inc. From wascott at sandia.gov Mon Aug 21 15:14:27 2017 From: wascott at sandia.gov (Scott, W Alan) Date: Mon, 21 Aug 2017 19:14:27 +0000 Subject: [Paraview-developers] Nice job on the ParaView downloads page Message-ID: <9f8c31385331493c9cddfb86c6c7bb09@ES01AMSNLNT.srn.sandia.gov> Nice job on the ParaView downloads page. Alan -------------------------------------------------------- W. Alan Scott ParaView Support Manager SAIC Sandia National Laboratories, MS 0807 Org 9326 - Building 880 A1-K (505) 284-0932 FAX (505) 284-5619 The most exciting phrase to hear in science is not "Eureka!" but "That's funny..." -- Isaac Asimov --------------------------------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From sysergie at gmail.com Tue Aug 22 08:15:31 2017 From: sysergie at gmail.com (Stanislav Sergienko) Date: Tue, 22 Aug 2017 07:15:31 -0500 Subject: [Paraview-developers] ParaView 5.4.0 PARAVIEW_BUILD_QT_GUI=ON QX11Info error Message-ID: I am trying to install ParaView 5.4.0 as part of OpenFOAM 5.0 on CentOS 7. I think I need to build it with QT Gui otherwise my PVReaders are failing to build. Here are my build options: ./makeParaView BUILD_SHARED_LIBS=ON \ VTK_USE_RPATH:BOOL=OFF \ PARAVIEW_BUILD_QT_GUI=ON \ VTK_USE_X=OFF \ OPENGL_INCLUDE_DIR=IGNORE \ OPENGL_xmesa_INCLUDE_DIR=IGNORE \ OPENGL_gl_LIBRARY=IGNORE \ OSMESA_INCLUDE_DIR=/usr/include \ OSMESA_LIBRARY=/usr/lib64/libOSMesa.so \ VTK_OPENGL_HAS_OSMESA=ON \ VTK_USE_OFFSCREEN=OFF \ CMAKE_BUILD_TYPE:STRING=Release \ PARAVIEW_ENABLE_PYTHON=ON \ PYTHON_INCLUDE_DIR=/soft/spack-0.10.1/opt/spack/linux- centos7-x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/include/python2.7 \ PYTHON_EXECUTABLE=/soft/spack-0.10.1/opt/spack/linux- centos7-x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/bin/python \ CMAKE_INSTALL_RPATH:STRING=/soft/spack-0.10.1/opt/spack/ linux-centos7-x86_64/gcc-7.1.0/ncurses-6.0-ezqafxmmhppan7mtqspbb56csmuoss fw/lib:/soft/spack-0.10.1/opt/spack/linux-centos7-x86_64/ gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/lib \ PARAVIEW_INSTALL_DEVELOPMENT_FILES=ON PYTHON_LIBRARY=/soft/spack-0. 10.1/opt/spack/linux-centos7-x86_64/gcc-7.1.0/python-2.7.13- qtrdup642rbar2ju33q7tffswry7hcun/lib/libpython2.7.so.1.0 -qmake /home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/bin/qmake \ CMAKE_PREFIX_PATH=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 \ Qt5_DIR=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 \ VTK_Group_Qt=ON And I am getting the following error. ParaView-5.4.0/VTK/GUISupport/Qt/QVTKWidget.cxx:50:11: fatal error: QX11Info: No such file or directory # include ^~~~~~~~~~ compilation terminated. make[2]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/QVTKWidget.cxx.o] Error 1 make[1]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/all] Error 2 I think my Qt 5.6.2 was built correctly and I see QX11Info headers and X11extras libararies are there. Any help is greatly appreciated. Thanks. Stas Sergienko -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Tue Aug 22 08:55:43 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Tue, 22 Aug 2017 08:55:43 -0400 Subject: [Paraview-developers] ParaView 5.4.0 PARAVIEW_BUILD_QT_GUI=ON QX11Info error In-Reply-To: References: Message-ID: Can you attach your CMakeCache.txt file please? Thanks Utkarsh On Tue, Aug 22, 2017 at 8:15 AM, Stanislav Sergienko wrote: > I am trying to install ParaView 5.4.0 as part of OpenFOAM 5.0 on CentOS 7. > I think I need to build it with QT Gui otherwise my PVReaders are failing > to build. > > Here are my build options: > > ./makeParaView BUILD_SHARED_LIBS=ON \ > VTK_USE_RPATH:BOOL=OFF \ > PARAVIEW_BUILD_QT_GUI=ON \ > VTK_USE_X=OFF \ > OPENGL_INCLUDE_DIR=IGNORE \ > OPENGL_xmesa_INCLUDE_DIR=IGNORE \ > OPENGL_gl_LIBRARY=IGNORE \ > OSMESA_INCLUDE_DIR=/usr/include \ > OSMESA_LIBRARY=/usr/lib64/libOSMesa.so \ > VTK_OPENGL_HAS_OSMESA=ON \ > VTK_USE_OFFSCREEN=OFF \ > CMAKE_BUILD_TYPE:STRING=Release \ > PARAVIEW_ENABLE_PYTHON=ON \ > PYTHON_INCLUDE_DIR=/soft/spack-0.10.1/opt/spack/linux-centos > 7-x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/include/python2.7 > \ > PYTHON_EXECUTABLE=/soft/spack-0.10.1/opt/spack/linux-centos7 > -x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/bin/python > \ > CMAKE_INSTALL_RPATH:STRING=/soft/spack-0.10.1/opt/spack/linu > x-centos7-x86_64/gcc-7.1.0/ncurses-6.0-ezqafxmmhppan7mtqs > pbb56csmuossfw/lib:/soft/spack-0.10.1/opt/spack/linux- > centos7-x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/lib > \ > PARAVIEW_INSTALL_DEVELOPMENT_FILES=ON PYTHON_LIBRARY=/soft/spack-0.1 > 0.1/opt/spack/linux-centos7-x86_64/gcc-7.1.0/python-2.7.13-q > trdup642rbar2ju33q7tffswry7hcun/lib/libpython2.7.so.1.0 -qmake > /home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/bin/qmake \ > CMAKE_PREFIX_PATH=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 > \ > Qt5_DIR=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 \ > VTK_Group_Qt=ON > > > And I am getting the following error. > > ParaView-5.4.0/VTK/GUISupport/Qt/QVTKWidget.cxx:50:11: fatal error: > QX11Info: No such file or directory > # include > ^~~~~~~~~~ > compilation terminated. > make[2]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/QVTKWidget.cxx.o] > Error 1 > make[1]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/all] Error > 2 > > I think my Qt 5.6.2 was built correctly and I see QX11Info headers and > X11extras libararies are there. > > Any help is greatly appreciated. > > Thanks. > > Stas Sergienko > > _______________________________________________ > Powered by www.kitware.com > > Visit other Kitware open-source projects at http://www.kitware.com/ > opensource/opensource.html > > Search the list archives at: http://markmail.org/search/?q= > Paraview-developers > > Follow this link to subscribe/unsubscribe: > http://public.kitware.com/mailman/listinfo/paraview-developers > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From utkarsh.ayachit at kitware.com Tue Aug 22 10:08:22 2017 From: utkarsh.ayachit at kitware.com (Utkarsh Ayachit) Date: Tue, 22 Aug 2017 10:08:22 -0400 Subject: [Paraview-developers] ParaView 5.4.0 PARAVIEW_BUILD_QT_GUI=ON QX11Info error In-Reply-To: References:

Message-ID: Stas, The issue is with VTK_USE_X being set to OFF. 5.4 cannot be build the Qt GUI when USE_X is off (ParaView "master" is a little more flexible there). Utkarsh On Tue, Aug 22, 2017 at 8:55 AM, Utkarsh Ayachit < utkarsh.ayachit at kitware.com> wrote: > Can you attach your CMakeCache.txt file please? > > Thanks > Utkarsh > > On Tue, Aug 22, 2017 at 8:15 AM, Stanislav Sergienko > wrote: > >> I am trying to install ParaView 5.4.0 as part of OpenFOAM 5.0 on CentOS >> 7. I think I need to build it with QT Gui otherwise my PVReaders are >> failing to build. >> >> Here are my build options: >> >> ./makeParaView BUILD_SHARED_LIBS=ON \ >> VTK_USE_RPATH:BOOL=OFF \ >> PARAVIEW_BUILD_QT_GUI=ON \ >> VTK_USE_X=OFF \ >> OPENGL_INCLUDE_DIR=IGNORE \ >> OPENGL_xmesa_INCLUDE_DIR=IGNORE \ >> OPENGL_gl_LIBRARY=IGNORE \ >> OSMESA_INCLUDE_DIR=/usr/include \ >> OSMESA_LIBRARY=/usr/lib64/libOSMesa.so \ >> VTK_OPENGL_HAS_OSMESA=ON \ >> VTK_USE_OFFSCREEN=OFF \ >> CMAKE_BUILD_TYPE:STRING=Release \ >> PARAVIEW_ENABLE_PYTHON=ON \ >> PYTHON_INCLUDE_DIR=/soft/spack-0.10.1/opt/spack/linux-centos >> 7-x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/include/python2.7 >> \ >> PYTHON_EXECUTABLE=/soft/spack-0.10.1/opt/spack/linux-centos7 >> -x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/bin/python >> \ >> CMAKE_INSTALL_RPATH:STRING=/soft/spack-0.10.1/opt/spack/linu >> x-centos7-x86_64/gcc-7.1.0/ncurses-6.0-ezqafxmmhppan7mtqspbb >> 56csmuossfw/lib:/soft/spack-0.10.1/opt/spack/linux-centos7- >> x86_64/gcc-7.1.0/python-2.7.13-qtrdup642rbar2ju33q7tffswry7hcun/lib \ >> PARAVIEW_INSTALL_DEVELOPMENT_FILES=ON PYTHON_LIBRARY=/soft/spack-0.1 >> 0.1/opt/spack/linux-centos7-x86_64/gcc-7.1.0/python-2.7.13-q >> trdup642rbar2ju33q7tffswry7hcun/lib/libpython2.7.so.1.0 -qmake >> /home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/bin/qmake \ >> CMAKE_PREFIX_PATH=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 >> \ >> Qt5_DIR=/home/ssergien/OpenFOAM/ThirdParty-5.0/Qt5/lib/cmake/Qt5 \ >> VTK_Group_Qt=ON >> >> >> And I am getting the following error. >> >> ParaView-5.4.0/VTK/GUISupport/Qt/QVTKWidget.cxx:50:11: fatal error: >> QX11Info: No such file or directory >> # include >> ^~~~~~~~~~ >> compilation terminated. >> make[2]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/QVTKWidget.cxx.o] >> Error 1 >> make[1]: *** [VTK/GUISupport/Qt/CMakeFiles/vtkGUISupportQt.dir/all] >> Error 2 >> >> I think my Qt 5.6.2 was built correctly and I see QX11Info headers and >> X11extras libararies are there. >> >> Any help is greatly appreciated. >> >> Thanks. >> >> Stas Sergienko >> >> _______________________________________________ >> Powered by www.kitware.com >> >> Visit other Kitware open-source projects at >> http://www.kitware.com/opensource/opensource.html >> >> Search the list archives at: http://markmail.org/search/?q= >> Paraview-developers >> >> Follow this link to subscribe/unsubscribe: >> http://public.kitware.com/mailman/listinfo/paraview-developers >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From cory.quammen at kitware.com Tue Aug 22 13:07:25 2017 From: cory.quammen at kitware.com (Cory Quammen) Date: Tue, 22 Aug 2017 13:07:25 -0400 Subject: [Paraview-developers] ParaView superbuild for 5.4.1 tagged Message-ID: Hi folks, ParaView's superbuild git repository has now been tagged with v5.4.1. Thanks, Cory -- Cory Quammen Staff R&D Engineer Kitware, Inc.