I got something working by initializing an attribute on the first frame where i pick certain points to keep based on their id and then i transfer that keep value to each frame of the simulation so it checks if the particle should stay or not. if(@id = point(1, "id", idtopoint(1, @id)) && point(1, "keep", idtopoint(1, @id)) == 1){ i@keep = 1; } Here the first input of the wrangle is the sim and the second is the frozen frame where i manually select points.

So i have a flip simulation where the water is in a lake and moving slightly up and down the land. I want to scatter some lilypads on it but am not sure how. I tried extracting points from the particle sim using their point id which kept the same points on each frame but ran into issues because i only want particles on the edge of the simulation not in the center of the lake so i was deleting all particles inside of a bounding object but particles were going in and out of this boundary causing newly appearing/disappearing points on some frames. Another problem i had is that not all the particles from the fluid sim are at the surface some are inside but i was able to somewhat get around this with a ray. Just looking for any advice, thankyou!

So im simming a waterfall into a lake but am running into problems with the whitewater sim. When i color the vorticity attribute on the fluid sim i get a beautiful result but when i go to place a whitewater source down it doesnt seem to respond accordingly. In the images i attached are from the same frame, you can see where the vorticity is colored i would get nice whitewater on the waterfall and some ripples and the hopefully where its black there would be no whitewater but when i go to place a whitewater source and run the simulation i get a completely different result. I also notice that changing the vorticity settings in the whitewatersource SOP do not give any changes when i change the range or remap values so i am not sure if i am missing something? thankyou.

So im using the uv layout sop to pack my primitives into a boundary, the problem i have is that im using the packing for printing so i need the objects to have no scaling so i set the scale to "Fixed". This was working fine until i was in a rush and now it has broken as you see in the picture i linked (https://imgur.com/a/5zd53vk). Any ideas on how to fix this? Thankyou.

Thanks! The thing is the more I looked at it was started thinking that its not a bug but its working how it is meant to, not totally sure but it seems makes sense that the reference would require a save path instead of writing into the main usd but the whole framework is just so confusing i don't really know.

So it seems the only way to procedurally expand all the variants is using the explore variants LOP. The problem I get when using this is that the file I am exploring variants on which Is brought in using a reference LOP gets an error of not having a save path and it defaults to creating a new usd file for the layer instead of reading the one from the reference LOP. I tried to look inside the network and noticed the reference lop there is what was causing the problems, but it was too complicated for me to fix. I tried just bypassing the reference LOP inside the network which worked but then I get the default prim name "/__variants" set in the python script and I was not able to make any changes to the primName variable without getting invalid null prim error. Please help! Edit: I realize after messing around a bit more that the reference LOP causes the separate USD file to be written out whereas other methods like extracting the variants manually or using a stage manager use write out the same data but inside the final usd layer from the usd rop. So basically I am just wondering how to procedurally get all variant prims to use for point instancing since the explore variants method that is often shown does not appear ideal.

I am calling an API and need to be able to display images that i am getting in a list, but i do not know how to do this. Not sure where to really start, i cannot find anything within the hou.ui module and i am unfamiliar with hou.qt and have not seen it used in any examples yet. The examples i have seen use the hutils.Qt module but this recent post says that there are no plans to upgrade Qt https://www.sidefx.com/forum/topic/82596/. Just looking to be put in the right direction, Thankyou!

I am calling an API and need to be able to display images that i am getting in a list, but i do not know how to do this. Not sure where to really start, i cannot find anything within the hou.ui module and i am unfamiliar with hou.qt and have not seen it used in any examples yet. The examples i have seen use the hutils.Qt module but this recent post says that there are no plans to upgrade Qt https://www.sidefx.com/forum/topic/82596/. Just looking to be put in the right direction, Thankyou!

I have installed rpr for houdini 19.383 the latest version and it matches my houdini version. I can get the render going in solaris but when i change any settings in the light there is no update even if i restart the render. I do get an error of "Device 0 not available (total number of devices -2018589406) at file C:\JN\WS\RadeonProImageProcessor_Build\src\RadeonImageFilters.cpp line 338" which is weird because thats not even a valid directory on my system. Materials seem to work well its just the lights, im running a gtx 1070.

That Fus.hip works very fast it is certainly an option i will look at. However i am still wondering if there is a way to use world coordinates with noise or something to get very smooth gradient like on the pighead below. I can then use the gradient to manipulate tons of values and it visualizes well and does not use a solver/recursion. If i were to do it with the iteration based method using pcfind yours is certainly faster but then for each iteration i would need to divide each expansion iteration by the iteration number to get a gradient and apply some randomness which kills a lot of the art direction you get with noise and ramps. ramps like here

Thankyou, but im looking for something not recursive and ideally not point calculation based. It seems junichiros method uses nearpoints and since im operating on a pointcloud i would have to use pcfind which is very slow.

Ill try to explain this my best, i have a point cloud with a group of points that are like patches on the pointcloud (the white points below). I want to start a gradient kind of like an infection outwards from these groups to end up creating a noise type look. Ideally in the end the white patches would have a float value of like 1 and the red points would have a float value of 0 with a gradient at their edge which would have a controllable length. Using a sort of infection type solver or pcfind functions has been very slow because i am starting from these point patches and i need it to be fairly fast and iterable. I was thinking some sort of world coordinate based system like a noise would work well for this since you dont do any point searching which would allow for easy gradients but i dont know how i can initiate it from these start patches. Thankyou!

I started off creating shelf tools, worked fine then I wanted to read modules on disk so i used the PYTHONPATH variable which worked great. However the second I go to run functions from my module I get the error 'module has no attribute' as shown below. import mymodule mymodule.my_function() it returns the error: module 'mymodule 'has no attribute 'myfunction'. My workaround right now is just symlinking my .py file into the /python3.7libs directory which reads fine for the moment.

I've been a bit confused about how you can animate lights in Solaris and USD overall. I assume you can put them inside an xform and animate manually but how would you go about procedurally animating them like if I wanted to attach them to a point or something?

Ah yes there was a weird name i found, clumsy mistake. Thanks for the function as well works perfectly!