[Paraview] Bug? - ExtractedSelection is mapped between Extracted Blocks
Hi, I understand that objects in a vtk pipeline will be updated if something earlier changes, but I can see no use for this behaviour (and it's quite a pain to me ;). To replicate: 1) Load the can.ex2 example and apply the 'Extract Block' filter. 2) Navigate to the ExtractBlock/'Block indices' menu and switch on 'Unnamed block ID 1' so you can see the semi-circular can object. 3) Click the 'Select Points On' rubber-band selection button and draw a rectangle of points on the can. 4) Click the 'Extract Selection' button. 5) Go back to the ExtractBlock/'Block indices' menu, switch off 'Unnamed block ID 1' and turn on 'Unnamed block ID 2' You will see that the extracted selection is mapped to apparently arbitrary points in or on the second block (you may need to change opacity to see the mapped points). Other similar but strange behaviour occurs if you select across blocks and switch to view just one or a few of them. I note that if I obtain the pqSelectionManager I can clear the selection, but as the docs say, the selection still exists somewhere in the vtk code. Is there a better way to control this action? (Tested with ParaView 4.1 and Windows and Ubuntu, and 4.2 on Ubuntu.) Thanks, Richard ps I tried searching through the archives for block/selection issues and I'm a bit surprised not to find anything like this before though it may be related to issue raised by Robert Ainsworth ?? [so I've just cc'd this to him] (http://paraview.markmail.org/message/z4eq4hj2zerrwste?q=selection+on+blockspage=3) ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
[Paraview] Labeling specific points in dataset
Hi Everyone, I have a dataset of about 1000 or so points and I want to label say every 100'th point with the point id. I can use the DataLabelRepresentation if I want to label every point, but as far as I can tell this cannot label only certain points it is everything or nothing. I could also create a new dataset with every 100th point in it and use that with the DataLabelRepresentation but then the point ids would be wrong. Any ideas or suggestions are welcome. Cheers Kit ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Labeling specific points in dataset
You can tell DataLabelRepresentation to label using any data array. By adding an array such as original ids or something during the extraction processes, you can tell the representation to use that array for labeling instead. That's pretty much on extract-selection filters work too. On Thu, Oct 16, 2014 at 7:29 AM, Kit Chambers kit.chambers...@gmail.com wrote: Hi Everyone, I have a dataset of about 1000 or so points and I want to label say every 100'th point with the point id. I can use the DataLabelRepresentation if I want to label every point, but as far as I can tell this cannot label only certain points it is everything or nothing. I could also create a new dataset with every 100th point in it and use that with the DataLabelRepresentation but then the point ids would be wrong. Any ideas or suggestions are welcome. Cheers Kit ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Catalyst question
Hi John, You are correct that the rendering that's done by Catalyst to create images is not directly available to the client. I hope to fix this in the not too distant future. The idea being that you select which Catalyst view you want to interact with and then it switches the view to that such that you can update view options like camera angles, lookup tables, etc. as well as output frequency and filename for that view when from Catalyst. As for creating a live connection to a Catalyst enabled simulation, maybe the thing you want is to use a live connection with pausing the simulation to get to data at a specific point in the simulation. If this data is too much to download to the client you can create a separate pvserver to connect to Catalyst and then connect your client to that pvserver. I haven't tried this but Utkarsh said that it was designed to work this way. Does this sound like a reasonable way of doing what you want? Now that the live functionality also has the ability to pause a simulation, it's feasible that we could go the libsim route and have Catalyst act as server for everything instead of downloading it to the client or separate server. We just haven't had a lot of impetus for providing that use case yet so we've focused on other stuff. Regards, Andy On Thu, Oct 16, 2014 at 8:58 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Andy I appear to be suffering from a confusion of what the catalyst framework does - or how it is used ... I had imagined that I’d run a simulation on N ranks, have catalyst bound to them and be able to do vis on those N ranks. But in fact, when I connect to a live simulation, I have to “extract” results from them and then render them back on the client. Is this the case or did I miss something? The parallel image generation appears to work as advertised, but the images go off to disk and not to the renderwindow. Is there a way on configuring it so that I can see the actual rendering taking place in the renderview? I wanted to display the contents of a custom render plugin using cuda based rendering on many GPUs and be able to interact with it, but I am obviously missing something in my setup. thanks JB *From:* Andy Bauer [mailto:andy.ba...@kitware.com] *Sent:* 26 August 2014 15:34 *To:* Biddiscombe, John A. *Cc:* paraview@paraview.org *Subject:* Re: [Paraview] Catalyst question Hi John, This live capability is now available and recently has been cleaned up quite a bit. There are still a couple more small changes that are waiting a code review process before they go into the master branch. The user's guide needs an update for a bunch of new information (live, catalyst editions, zero-copy arrays) that's important for Catalyst. We're also doing some preliminary testing with a newly developed ADIOS writer for better IO performance as well. Getting back to the live viz stuff, when you create a Catalyst Python script and export it using the wizard you'll have the option to enable live viz (it's the same wizard page as the screenshot output specification). Then, while your simulation is running use the Catalyst-Connect option to connect. There are two tests (pv.CatalystLiveSetBreakpoint and pvcs.CatalystLiveSetBreakpoint) that demonstrate the newly implemented live functionality which now includes the option to set a breakpoint in the simulation. If you get a chance to try it out, let me know what you think. Andy On Tue, Aug 26, 2014 at 3:40 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Dear Catalyst people I dragged out an old simulation demo I created some time ago and got it working with paraview/catalyst from git master branch - my plan was to test a live visualization of the simulation with the paraview GUI connected to the simulation. I had thought this capability was now possible, but I have not been following development closely and it looks like I was mistaken. Is it possible to visualize the simulation live - rather than just generate images (which works fine). (I had a quick skim through the user manual, but didn’t see this feature mentioned). Thanks JB -- John Biddiscombe,email:biddisco @.at.@ cscs.ch http://www.cscs.ch/ CSCS, Swiss National Supercomputing Centre | Tel: +41 (91) 610.82.07 Via Trevano 131, 6900 Lugano, Switzerland | Fax: +41 (91) 610.82.82 ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep
Re: [Paraview] Catalyst question
Hi John, I had imagined that I’d run a simulation on N ranks, have catalyst bound to them and be able to do vis on those N ranks. But in fact, when I connect to a live simulation, I have to “extract” results from them and then render them back on the client. Is this the case or did I miss something? Just to confirm Andy's answer, you can do the rendering on a separate ParaView server (pvserver), which can be the same as the server you run the simulation on. Catalyst will send extracts to this server which will render them. Only images will be send back to the client. We show that use case at the end of our article. http://www.kitware.com/blog/home/post/722 Dan The parallel image generation appears to work as advertised, but the images go off to disk and not to the renderwindow. Is there a way on configuring it so that I can see the actual rendering taking place in the renderview? I wanted to display the contents of a custom render plugin using cuda based rendering on many GPUs and be able to interact with it, but I am obviously missing something in my setup. thanks JB *From:* Andy Bauer [mailto:andy.ba...@kitware.com] *Sent:* 26 August 2014 15:34 *To:* Biddiscombe, John A. *Cc:* paraview@paraview.org *Subject:* Re: [Paraview] Catalyst question Hi John, This live capability is now available and recently has been cleaned up quite a bit. There are still a couple more small changes that are waiting a code review process before they go into the master branch. The user's guide needs an update for a bunch of new information (live, catalyst editions, zero-copy arrays) that's important for Catalyst. We're also doing some preliminary testing with a newly developed ADIOS writer for better IO performance as well. Getting back to the live viz stuff, when you create a Catalyst Python script and export it using the wizard you'll have the option to enable live viz (it's the same wizard page as the screenshot output specification). Then, while your simulation is running use the Catalyst-Connect option to connect. There are two tests (pv.CatalystLiveSetBreakpoint and pvcs.CatalystLiveSetBreakpoint) that demonstrate the newly implemented live functionality which now includes the option to set a breakpoint in the simulation. If you get a chance to try it out, let me know what you think. Andy On Tue, Aug 26, 2014 at 3:40 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Dear Catalyst people I dragged out an old simulation demo I created some time ago and got it working with paraview/catalyst from git master branch - my plan was to test a live visualization of the simulation with the paraview GUI connected to the simulation. I had thought this capability was now possible, but I have not been following development closely and it looks like I was mistaken. Is it possible to visualize the simulation live - rather than just generate images (which works fine). (I had a quick skim through the user manual, but didn’t see this feature mentioned). Thanks JB -- John Biddiscombe,email:biddisco @.at.@ cscs.ch http://www.cscs.ch/ CSCS, Swiss National Supercomputing Centre | Tel: +41 (91) 610.82.07 Via Trevano 131, 6900 Lugano, Switzerland | Fax: +41 (91) 610.82.82 ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Catalyst question
Andy, Dan Thank you both for clarifying this. I might give the catalyst-server-client version a go, but it seems like it might be too much trouble ... I expect I’ll just use the solution we developed here at CSCS a couple of years ago where we just connect the simulation to the pvservers and then connect the client to them. That way I won’t need to copy the data between processes again (there is already a copy because the simulation is writing to a shared memory buffer and the pvservers would read from that - not the same as the H5FDdsm stuff we developed, but very similar, and I can simply piggy back our live viz on top of that). We wanted to do a live demo of GPU rendering on 1024 or so nodes at SC14 so it may be that using something I’m already familiar with will save me time as were getting close to Nov ... On that note : It ought to be trivial to whip up a catalyst “view” which captures the images from the internal renderwindow created by catalyst and have them displayed in the GUI so that the user can see the ‘live’ viz. Not sure how you’d pass the interactions back to handle mouse etc, but if the client can write to the camera objects via proxies then it should be possible might try this myself if I have a moment spare. thanks again JB From: Dan Lipsa [mailto:dan.li...@kitware.com] Sent: 16 October 2014 16:32 To: Biddiscombe, John A. Cc: Andy Bauer; paraview@paraview.org Subject: Re: [Paraview] Catalyst question Hi John, I had imagined that I’d run a simulation on N ranks, have catalyst bound to them and be able to do vis on those N ranks. But in fact, when I connect to a live simulation, I have to “extract” results from them and then render them back on the client. Is this the case or did I miss something? Just to confirm Andy's answer, you can do the rendering on a separate ParaView server (pvserver), which can be the same as the server you run the simulation on. Catalyst will send extracts to this server which will render them. Only images will be send back to the client. We show that use case at the end of our article. http://www.kitware.com/blog/home/post/722 Dan The parallel image generation appears to work as advertised, but the images go off to disk and not to the renderwindow. Is there a way on configuring it so that I can see the actual rendering taking place in the renderview? I wanted to display the contents of a custom render plugin using cuda based rendering on many GPUs and be able to interact with it, but I am obviously missing something in my setup. thanks JB From: Andy Bauer [mailto:andy.ba...@kitware.commailto:andy.ba...@kitware.com] Sent: 26 August 2014 15:34 To: Biddiscombe, John A. Cc: paraview@paraview.orgmailto:paraview@paraview.org Subject: Re: [Paraview] Catalyst question Hi John, This live capability is now available and recently has been cleaned up quite a bit. There are still a couple more small changes that are waiting a code review process before they go into the master branch. The user's guide needs an update for a bunch of new information (live, catalyst editions, zero-copy arrays) that's important for Catalyst. We're also doing some preliminary testing with a newly developed ADIOS writer for better IO performance as well. Getting back to the live viz stuff, when you create a Catalyst Python script and export it using the wizard you'll have the option to enable live viz (it's the same wizard page as the screenshot output specification). Then, while your simulation is running use the Catalyst-Connect option to connect. There are two tests (pv.CatalystLiveSetBreakpoint and pvcs.CatalystLiveSetBreakpoint) that demonstrate the newly implemented live functionality which now includes the option to set a breakpoint in the simulation. If you get a chance to try it out, let me know what you think. Andy On Tue, Aug 26, 2014 at 3:40 AM, Biddiscombe, John A. biddi...@cscs.chmailto:biddi...@cscs.ch wrote: Dear Catalyst people I dragged out an old simulation demo I created some time ago and got it working with paraview/catalyst from git master branch - my plan was to test a live visualization of the simulation with the paraview GUI connected to the simulation. I had thought this capability was now possible, but I have not been following development closely and it looks like I was mistaken. Is it possible to visualize the simulation live - rather than just generate images (which works fine). (I had a quick skim through the user manual, but didn’t see this feature mentioned). Thanks JB -- John Biddiscombe,email:biddisco @.at.@ cscs.chhttp://cscs.ch http://www.cscs.ch/ CSCS, Swiss National Supercomputing Centre | Tel: +41 (91) 610.82.07tel:%2B41%20%2891%29%20610.82.07 Via Trevano 131, 6900 Lugano, Switzerland | Fax: +41 (91) 610.82.82tel:%2B41%20%2891%29%20610.82.82 ___ Powered by www.kitware.comhttp://www.kitware.com Visit
Re: [Paraview] Catalyst question
Hi Dan, Thanks for the correction -- I should have read your post closer! Thanks, Andy On Thu, Oct 16, 2014 at 10:32 AM, Dan Lipsa dan.li...@kitware.com wrote: Hi John, I had imagined that I’d run a simulation on N ranks, have catalyst bound to them and be able to do vis on those N ranks. But in fact, when I connect to a live simulation, I have to “extract” results from them and then render them back on the client. Is this the case or did I miss something? Just to confirm Andy's answer, you can do the rendering on a separate ParaView server (pvserver), which can be the same as the server you run the simulation on. Catalyst will send extracts to this server which will render them. Only images will be send back to the client. We show that use case at the end of our article. http://www.kitware.com/blog/home/post/722 Dan The parallel image generation appears to work as advertised, but the images go off to disk and not to the renderwindow. Is there a way on configuring it so that I can see the actual rendering taking place in the renderview? I wanted to display the contents of a custom render plugin using cuda based rendering on many GPUs and be able to interact with it, but I am obviously missing something in my setup. thanks JB *From:* Andy Bauer [mailto:andy.ba...@kitware.com] *Sent:* 26 August 2014 15:34 *To:* Biddiscombe, John A. *Cc:* paraview@paraview.org *Subject:* Re: [Paraview] Catalyst question Hi John, This live capability is now available and recently has been cleaned up quite a bit. There are still a couple more small changes that are waiting a code review process before they go into the master branch. The user's guide needs an update for a bunch of new information (live, catalyst editions, zero-copy arrays) that's important for Catalyst. We're also doing some preliminary testing with a newly developed ADIOS writer for better IO performance as well. Getting back to the live viz stuff, when you create a Catalyst Python script and export it using the wizard you'll have the option to enable live viz (it's the same wizard page as the screenshot output specification). Then, while your simulation is running use the Catalyst-Connect option to connect. There are two tests (pv.CatalystLiveSetBreakpoint and pvcs.CatalystLiveSetBreakpoint) that demonstrate the newly implemented live functionality which now includes the option to set a breakpoint in the simulation. If you get a chance to try it out, let me know what you think. Andy On Tue, Aug 26, 2014 at 3:40 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Dear Catalyst people I dragged out an old simulation demo I created some time ago and got it working with paraview/catalyst from git master branch - my plan was to test a live visualization of the simulation with the paraview GUI connected to the simulation. I had thought this capability was now possible, but I have not been following development closely and it looks like I was mistaken. Is it possible to visualize the simulation live - rather than just generate images (which works fine). (I had a quick skim through the user manual, but didn’t see this feature mentioned). Thanks JB -- John Biddiscombe,email:biddisco @.at.@ cscs.ch http://www.cscs.ch/ CSCS, Swiss National Supercomputing Centre | Tel: +41 (91) 610.82.07 Via Trevano 131, 6900 Lugano, Switzerland | Fax: +41 (91) 610.82.82 ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Catalyst question
Hi John, I'll be at SC14 and I'd be interested in seeing your demo. Are the details of when and where ready yet? Regards, Andy On Thu, Oct 16, 2014 at 10:43 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Andy, Dan Thank you both for clarifying this. I might give the catalyst-server-client version a go, but it seems like it might be too much trouble ... I expect I’ll just use the solution we developed here at CSCS a couple of years ago where we just connect the simulation to the pvservers and then connect the client to them. That way I won’t need to copy the data between processes again (there is already a copy because the simulation is writing to a shared memory buffer and the pvservers would read from that - not the same as the H5FDdsm stuff we developed, but very similar, and I can simply piggy back our live viz on top of that). We wanted to do a live demo of GPU rendering on 1024 or so nodes at SC14 so it may be that using something I’m already familiar with will save me time as were getting close to Nov ... On that note : It ought to be trivial to whip up a catalyst “view” which captures the images from the internal renderwindow created by catalyst and have them displayed in the GUI so that the user can see the ‘live’ viz. Not sure how you’d pass the interactions back to handle mouse etc, but if the client can write to the camera objects via proxies then it should be possible might try this myself if I have a moment spare. thanks again JB *From:* Dan Lipsa [mailto:dan.li...@kitware.com] *Sent:* 16 October 2014 16:32 *To:* Biddiscombe, John A. *Cc:* Andy Bauer; paraview@paraview.org *Subject:* Re: [Paraview] Catalyst question Hi John, I had imagined that I’d run a simulation on N ranks, have catalyst bound to them and be able to do vis on those N ranks. But in fact, when I connect to a live simulation, I have to “extract” results from them and then render them back on the client. Is this the case or did I miss something? Just to confirm Andy's answer, you can do the rendering on a separate ParaView server (pvserver), which can be the same as the server you run the simulation on. Catalyst will send extracts to this server which will render them. Only images will be send back to the client. We show that use case at the end of our article. http://www.kitware.com/blog/home/post/722 Dan The parallel image generation appears to work as advertised, but the images go off to disk and not to the renderwindow. Is there a way on configuring it so that I can see the actual rendering taking place in the renderview? I wanted to display the contents of a custom render plugin using cuda based rendering on many GPUs and be able to interact with it, but I am obviously missing something in my setup. thanks JB *From:* Andy Bauer [mailto:andy.ba...@kitware.com] *Sent:* 26 August 2014 15:34 *To:* Biddiscombe, John A. *Cc:* paraview@paraview.org *Subject:* Re: [Paraview] Catalyst question Hi John, This live capability is now available and recently has been cleaned up quite a bit. There are still a couple more small changes that are waiting a code review process before they go into the master branch. The user's guide needs an update for a bunch of new information (live, catalyst editions, zero-copy arrays) that's important for Catalyst. We're also doing some preliminary testing with a newly developed ADIOS writer for better IO performance as well. Getting back to the live viz stuff, when you create a Catalyst Python script and export it using the wizard you'll have the option to enable live viz (it's the same wizard page as the screenshot output specification). Then, while your simulation is running use the Catalyst-Connect option to connect. There are two tests (pv.CatalystLiveSetBreakpoint and pvcs.CatalystLiveSetBreakpoint) that demonstrate the newly implemented live functionality which now includes the option to set a breakpoint in the simulation. If you get a chance to try it out, let me know what you think. Andy On Tue, Aug 26, 2014 at 3:40 AM, Biddiscombe, John A. biddi...@cscs.ch wrote: Dear Catalyst people I dragged out an old simulation demo I created some time ago and got it working with paraview/catalyst from git master branch - my plan was to test a live visualization of the simulation with the paraview GUI connected to the simulation. I had thought this capability was now possible, but I have not been following development closely and it looks like I was mistaken. Is it possible to visualize the simulation live - rather than just generate images (which works fine). (I had a quick skim through the user manual, but didn’t see this feature mentioned). Thanks JB -- John Biddiscombe,email:biddisco @.at.@ cscs.ch http://www.cscs.ch/ CSCS, Swiss National Supercomputing Centre | Tel: +41 (91) 610.82.07
Re: [Paraview] OpenGL2 backend in 4.2
Greg, This isn't ready for testing yet. There are still code changes that need to go into ParaView before folks can start building with OpenGL2 backend support. Utkarsh On Thu, Oct 16, 2014 at 11:26 AM, Greg Abram g...@tacc.utexas.edu wrote: Hey y'all - Anyone able to get the OpenGL2 backend to build? If I set VTK_RENDERING_BACKEND to OpenGL2, I get: CMake Error at VTK/Rendering/OpenGL/CMakeLists.txt:3 (message): vtkRenderingOpenGL cannot be built with vtkRenderingOpenGL2, please disable one of them. And I can't figure out how to get past it. Greg Gregory D. Abram, Ph.D. Research Engineering/Scientist Associate Texas Advanced Computing Center The University of Texas at Austin (512) 471-8196 g...@tacc.utexas.edu [image: TACC Website] https://www.tacc.utexas.edu/ Connect With TACC https://www.tacc.utexas.edu/ ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] Bug? - ExtractedSelection is mapped between Extracted Blocks
Try unchecking the Purne Output checkbox on the Extract Block filter's Property panel. The problem is that Prune Output changes the structure of the multiblock. Hence the selection defined in (3) resolves differently after (5). Utkarsh On Thu, Oct 16, 2014 at 7:03 AM, redkite redk...@virgin.net wrote: Hi, I understand that objects in a vtk pipeline will be updated if something earlier changes, but I can see no use for this behaviour (and it's quite a pain to me ;). To replicate: 1) Load the can.ex2 example and apply the 'Extract Block' filter. 2) Navigate to the ExtractBlock/'Block indices' menu and switch on 'Unnamed block ID 1' so you can see the semi-circular can object. 3) Click the 'Select Points On' rubber-band selection button and draw a rectangle of points on the can. 4) Click the 'Extract Selection' button. 5) Go back to the ExtractBlock/'Block indices' menu, switch off 'Unnamed block ID 1' and turn on 'Unnamed block ID 2' You will see that the extracted selection is mapped to apparently arbitrary points in or on the second block (you may need to change opacity to see the mapped points). Other similar but strange behaviour occurs if you select across blocks and switch to view just one or a few of them. I note that if I obtain the pqSelectionManager I can clear the selection, but as the docs say, the selection still exists somewhere in the vtk code. Is there a better way to control this action? (Tested with ParaView 4.1 and Windows and Ubuntu, and 4.2 on Ubuntu.) Thanks, Richard ps I tried searching through the archives for block/selection issues and I'm a bit surprised not to find anything like this before though it may be related to issue raised by Robert Ainsworth ?? [so I've just cc'd this to him] (http://paraview.markmail.org/message/z4eq4hj2zerrwste?q=selection+on+blockspage=3) ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] using calculator filter on a table?
I think I've tracked down the problem...I'll have a fix soon. Utkarsh On Thu, Sep 25, 2014 at 6:18 PM, Joshua Murphy joshua.mur...@lasp.colorado.edu wrote: Hi Utkarsh, I have tried this once again in v4.2.0, and I am getting the same segfault. any thoughts? my input is a multi-block table... -Josh From: Utkarsh Ayachit [utkarsh.ayac...@kitware.com] Sent: Thursday, September 25, 2014 11:23 AM To: Joshua Murphy Cc: paraview@paraview.org Subject: Re: [Paraview] using calculator filter on a table? For now, with 4.2 RC1, stick to using Programmable Filter with Script as follows: expr = inputs[0].RowData[RTData] * 2.0 output.RowData.append(expr, result) If you can add a bug to the bug tracker for this, that'd be great too. Thanks Utkarsh On Thu, Sep 25, 2014 at 12:45 PM, Joshua Murphy joshua.mur...@lasp.colorado.edu wrote: Hello all, Is there a way to use the Calculator filter on a table? My reader loads a bunch of auxiliary files that loads time-series information for a specific point (at a higher temporal resolution than the 3D model output). I need to be able to convert the units in the time series information with some simple calculations so that I can match the units of the remainder of my work. Is there a way to use the calculator filter to do this? It is greyed out when I have the table selected... Are there some other steps I need to do to make it work? Thanks, Josh ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview ___ 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 Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/paraview
Re: [Paraview] BOX - volume?
Hi Bertwim, The problem you're experiencing is not an issue of missing extensions, but rather a bug in ATI's driver. Projected tetrahedra is disabled when an ATI device is detected to avoid that bug(ParaView will hang). However it's not a big deal. because you can use the Bunyk raycast algrothim instead. After your data is on screen scroll to the bottom of the properties panel and look for the field Select mapper, set that to Bunyk ray cast, then change the representation to volume. Burlen On 10/15/2014 06:58 PM, B.W.H. van Beest wrote: I am still busy with finding out which etensions are needed, and how I can enable them. THis opebnGL-world is a bit alien to me, so it will take some time. Regards, Bertwim On 10/14/2014 11:07 PM, Aashish Chaudhary wrote: Looks like certain extensions are still not quite working on yours system even with the proprietary driver. I will see if I can run your code on my laptop but that would be for tomorrow. ** Assuming that this is not data dependent issue (which it should not). I think its probably this line causing the issue: bool driverSupport #if defined(__APPLE__) = true; #else = !(extensions-DriverIsATI() (extensions-GetDriverGLVersionMajor() = 3)) || extensions-GetIgnoreDriverBugs(ATI texturing bug); #endif On line number 181: can you change this code to : this-HasHardwareSupport=renwin!=NULLthis-IsSupported(renwin); to this-HasHardwareSupport=1; - Aashish On Tue, Oct 14, 2014 at 4:53 PM, B.W.H. van Beest b...@xs4all.nl mailto:b...@xs4all.nl wrote: Progress indeed. But I don't understand what you are saying about the transfer function. I' starting to wonder if I have two issues occurring simultaneously. In a new thread I just posted I described my problem in getting the colours in place. I need to write to a vtk file first and load it back. I just found that when I just did that and then switched fo volume rendering, an error was there about the absence of an opengl extension: ERROR: In /usr/local/git/ParaView/VTK/Rendering/VolumeOpenGL/vtkOpenGLProjectedTetrahedraMapper.cxx, line 185 vtkOpenGLProjectedTetrahedraMapper (0x58ae490): The required extensions are not supported. This is complete unknown terrain for me. Regards, Bertwim On 10/14/2014 09:58 PM, Burlen Loring wrote: progress! it sounds like volume rendering is working on the wavelet dataset. You do need the scalars for volume rendering, so make sure you have those when you try. You need to select the scalars in the drop down also, I usually do this before changing to the volume representation. You also may need to edit the transfer function. For testing you could make the opacity over the entire scalar range in which case you should see a solid cube. If you've done all that and don't see anything it's likely that the dataset is messed up. Add a dataset-Print(cerr) statement to the end of your request data, look in the information panel and verify correct number of points and scalar ranges. If all else fails post the dataset to the list. After creating your cube go to the file menu and save data. On 10/14/2014 10:40 AM, B.W.H. van Beest wrote: Hi Burlen, I have upgraded to the AMD drivers. Thanks for the hint. It has improved the situation a bit. Furthermore, I created the wavelet source. When selecting volume representation, I do see a three 3 picture, in all 5 volume rendering mode (smart, raycast and texture mapping, ray cast only, texture mapping only, GPU based). The picture a rather fuzzy, but then I don't know what I'm supposed to see here. The good news is that the openGL errors I had before when displaying a wavelet, now have gone. The situation with my own cube, though, has unchanged. unfortunately. The cube disappears in volume representation node, also when I re-inserted scalar values for each of the 8 vertices. Regards, Bertwim On 10/14/2014 06:06 PM, Burlen Loring wrote: Hi Bertwim, In recent releases of ParaView I have experienced issues/flakeyness with volume rendering. Something you should try: Create the wavelet source, then in the properties panel find volume rendering mode and set it to ray cast only. This has worked for me on a number of systems. may or may not work, but worth a try. The output you pasted below shows that although you've got a sweet ATI GPU you're not taking advantage of it! You're using the Intel integrated GPU with Mesa drivers. If I were you I would install proprietary the AMD drivers http://support.amd.com/en-us/download. This should give you nice performance boost and may even solve the OpenGL errors you reported, which these days are as often (or more) a result of bugs in the Mesa
Re: [Paraview] BOX - volume?
Got it! Impressive, your knowledge on these matters. Regards, Bertwim On 10/16/2014 07:53 PM, Burlen Loring wrote: Hi Bertwim, The problem you're experiencing is not an issue of missing extensions, but rather a bug in ATI's driver. Projected tetrahedra is disabled when an ATI device is detected to avoid that bug(ParaView will hang). However it's not a big deal. because you can use the Bunyk raycast algrothim instead. After your data is on screen scroll to the bottom of the properties panel and look for the field Select mapper, set that to Bunyk ray cast, then change the representation to volume. Burlen On 10/15/2014 06:58 PM, B.W.H. van Beest wrote: I am still busy with finding out which etensions are needed, and how I can enable them. THis opebnGL-world is a bit alien to me, so it will take some time. Regards, Bertwim On 10/14/2014 11:07 PM, Aashish Chaudhary wrote: Looks like certain extensions are still not quite working on yours system even with the proprietary driver. I will see if I can run your code on my laptop but that would be for tomorrow. ** Assuming that this is not data dependent issue (which it should not). I think its probably this line causing the issue: bool driverSupport #if defined(__APPLE__) = true; #else = !(extensions-DriverIsATI() (extensions-GetDriverGLVersionMajor() = 3)) || extensions-GetIgnoreDriverBugs(ATI texturing bug); #endif On line number 181: can you change this code to : this-HasHardwareSupport=renwin!=NULLthis-IsSupported(renwin); to this-HasHardwareSupport=1; - Aashish On Tue, Oct 14, 2014 at 4:53 PM, B.W.H. van Beest b...@xs4all.nl mailto:b...@xs4all.nl wrote: Progress indeed. But I don't understand what you are saying about the transfer function. I' starting to wonder if I have two issues occurring simultaneously. In a new thread I just posted I described my problem in getting the colours in place. I need to write to a vtk file first and load it back. I just found that when I just did that and then switched fo volume rendering, an error was there about the absence of an opengl extension: ERROR: In /usr/local/git/ParaView/VTK/Rendering/VolumeOpenGL/vtkOpenGLProjectedTetrahedraMapper.cxx, line 185 vtkOpenGLProjectedTetrahedraMapper (0x58ae490): The required extensions are not supported. This is complete unknown terrain for me. Regards, Bertwim On 10/14/2014 09:58 PM, Burlen Loring wrote: progress! it sounds like volume rendering is working on the wavelet dataset. You do need the scalars for volume rendering, so make sure you have those when you try. You need to select the scalars in the drop down also, I usually do this before changing to the volume representation. You also may need to edit the transfer function. For testing you could make the opacity over the entire scalar range in which case you should see a solid cube. If you've done all that and don't see anything it's likely that the dataset is messed up. Add a dataset-Print(cerr) statement to the end of your request data, look in the information panel and verify correct number of points and scalar ranges. If all else fails post the dataset to the list. After creating your cube go to the file menu and save data. On 10/14/2014 10:40 AM, B.W.H. van Beest wrote: Hi Burlen, I have upgraded to the AMD drivers. Thanks for the hint. It has improved the situation a bit. Furthermore, I created the wavelet source. When selecting volume representation, I do see a three 3 picture, in all 5 volume rendering mode (smart, raycast and texture mapping, ray cast only, texture mapping only, GPU based). The picture a rather fuzzy, but then I don't know what I'm supposed to see here. The good news is that the openGL errors I had before when displaying a wavelet, now have gone. The situation with my own cube, though, has unchanged. unfortunately. The cube disappears in volume representation node, also when I re-inserted scalar values for each of the 8 vertices. Regards, Bertwim On 10/14/2014 06:06 PM, Burlen Loring wrote: Hi Bertwim, In recent releases of ParaView I have experienced issues/flakeyness with volume rendering. Something you should try: Create the wavelet source, then in the properties panel find volume rendering mode and set it to ray cast only. This has worked for me on a number of systems. may or may not work, but worth a try. The output you pasted below shows that although you've got a sweet ATI GPU you're not taking advantage of it! You're using the Intel integrated GPU with Mesa drivers. If I were you I would install proprietary the AMD drivers
Re: [Paraview] BOX - volume?
Got it! Impressive, your knowledge on these matters ... Thanks Regards, Bertwim On 10/16/2014 07:53 PM, Burlen Loring wrote: Hi Bertwim, The problem you're experiencing is not an issue of missing extensions, but rather a bug in ATI's driver. Projected tetrahedra is disabled when an ATI device is detected to avoid that bug(ParaView will hang). However it's not a big deal. because you can use the Bunyk raycast algrothim instead. After your data is on screen scroll to the bottom of the properties panel and look for the field Select mapper, set that to Bunyk ray cast, then change the representation to volume. Burlen On 10/15/2014 06:58 PM, B.W.H. van Beest wrote: I am still busy with finding out which etensions are needed, and how I can enable them. THis opebnGL-world is a bit alien to me, so it will take some time. Regards, Bertwim On 10/14/2014 11:07 PM, Aashish Chaudhary wrote: Looks like certain extensions are still not quite working on yours system even with the proprietary driver. I will see if I can run your code on my laptop but that would be for tomorrow. ** Assuming that this is not data dependent issue (which it should not). I think its probably this line causing the issue: bool driverSupport #if defined(__APPLE__) = true; #else = !(extensions-DriverIsATI() (extensions-GetDriverGLVersionMajor() = 3)) || extensions-GetIgnoreDriverBugs(ATI texturing bug); #endif On line number 181: can you change this code to : this-HasHardwareSupport=renwin!=NULLthis-IsSupported(renwin); to this-HasHardwareSupport=1; - Aashish On Tue, Oct 14, 2014 at 4:53 PM, B.W.H. van Beest b...@xs4all.nl mailto:b...@xs4all.nl wrote: Progress indeed. But I don't understand what you are saying about the transfer function. I' starting to wonder if I have two issues occurring simultaneously. In a new thread I just posted I described my problem in getting the colours in place. I need to write to a vtk file first and load it back. I just found that when I just did that and then switched fo volume rendering, an error was there about the absence of an opengl extension: ERROR: In /usr/local/git/ParaView/VTK/Rendering/VolumeOpenGL/vtkOpenGLProjectedTetrahedraMapper.cxx, line 185 vtkOpenGLProjectedTetrahedraMapper (0x58ae490): The required extensions are not supported. This is complete unknown terrain for me. Regards, Bertwim On 10/14/2014 09:58 PM, Burlen Loring wrote: progress! it sounds like volume rendering is working on the wavelet dataset. You do need the scalars for volume rendering, so make sure you have those when you try. You need to select the scalars in the drop down also, I usually do this before changing to the volume representation. You also may need to edit the transfer function. For testing you could make the opacity over the entire scalar range in which case you should see a solid cube. If you've done all that and don't see anything it's likely that the dataset is messed up. Add a dataset-Print(cerr) statement to the end of your request data, look in the information panel and verify correct number of points and scalar ranges. If all else fails post the dataset to the list. After creating your cube go to the file menu and save data. On 10/14/2014 10:40 AM, B.W.H. van Beest wrote: Hi Burlen, I have upgraded to the AMD drivers. Thanks for the hint. It has improved the situation a bit. Furthermore, I created the wavelet source. When selecting volume representation, I do see a three 3 picture, in all 5 volume rendering mode (smart, raycast and texture mapping, ray cast only, texture mapping only, GPU based). The picture a rather fuzzy, but then I don't know what I'm supposed to see here. The good news is that the openGL errors I had before when displaying a wavelet, now have gone. The situation with my own cube, though, has unchanged. unfortunately. The cube disappears in volume representation node, also when I re-inserted scalar values for each of the 8 vertices. Regards, Bertwim On 10/14/2014 06:06 PM, Burlen Loring wrote: Hi Bertwim, In recent releases of ParaView I have experienced issues/flakeyness with volume rendering. Something you should try: Create the wavelet source, then in the properties panel find volume rendering mode and set it to ray cast only. This has worked for me on a number of systems. may or may not work, but worth a try. The output you pasted below shows that although you've got a sweet ATI GPU you're not taking advantage of it! You're using the Intel integrated GPU with Mesa drivers. If I were you I would install proprietary the AMD drivers