Integrate Custom Schemas

How to integrate a custom USD schema like a native Houdini primitive node.

Native Node

Let’s take a look at the native Render Settings primitive node:

It contains a couple of important elements:

• A Frame Range Sampling menu (useful to get rid of time dependencies)
• The Primitive Path parm (if in Create Mode)
• The Primitives parm (if in Edit mode)
• The Action menu, to either Create or Edit the primitive
• The Initialize Parameters menu, which lets you batch control the USD parms or retrieve another primitive attribute values (if in Edit mode)
• The Create Primitives menu, which allows you to control how and where the primitive will be created (if in Create mode)

Custom Node (HDA)

In order to emulate all those, we can create an HDA with the following structure:

Internal Nodes

Error

The Error node contains the following code on the Report this Error parm:

node = hou.pwd()
node_parent = node.parent()
inputs = node.inputs()
if inputs and node_parent.evalParm("createprims") == 1:
    prim = node_parent.evalParm("primpath")
    stage = inputs[0].stage()
    return int(stage.GetPrimAtPath(prim).IsValid())
return 0

It will return this error to the user:

An existing `chs("hda_label")` prim has been found at `chs("../primpath")`. You can only have one in your current scene. For selective edits, use a `chs("hda_label")` Edit instead.

if another primitive with the same path is found, and you’re about to override it by creating a new one on the same path .

hda_label is a spare parameter created on the Error node with a Python expression returning the value Fxquinox Context Info :

parent_node = hou.pwd().parent()
parent_type = parent_node.type()
parent_hda_definition = parent_type.definition()
author = parent_type.nameComponents()[1].title()
description = parent_hda_definition.description()
return f"{author} {description}"

Edit Properties from Node

This node is where the magic happens. It’s pretty much the same as the Edit Properties node, with a nice change:

That allows us to add the properties on the HDA itself: they will always be read from it and added accordingly.

This node is also linked to the Frame Range Sampling menu, Primitive Path parm, Primitives parm, Action menu, and the Initialize Parameters menu. It’s the one that allows our HDA to behave just like a native node.

Promote USD Attributes

In order for the Edit Properties from Node node to work, we need to bring the USD schema attributes to the HDA interface. With a custom schema, this gets very easy. Open the HDA Type Properties , and inside the Parameters tab navigate to the From USD tab. You can now filter the schema you want and add all its attributes to your parameters.

They will be automatically read by the Edit Properties from Node node!

Add an Edit Alias

On most of the native nodes creating a primitive, you’ll get two options: the node, and the node edit . Here's what gets displayed when pressing TAB inside the Network View to drop a Render Settings node:

You have the option to use a regular Render Settings , or a Render Settings Edit . Under the hood, this is a simple tool alias, which sets the Action menu to Edit , and sets all USD parameters to none , effectively bypassing any values that could have been added on the node parameters.

Here’s how it’s done for Render Settings (navigate to Render Settings > Type Properties > Interactive > Shelf Tools )

Inside the Options tab:

And inside the Script tab:

On our custom HDA Type Properties window, inside the Interactive > Shelf Tools tab, we can click on Create New and navigate to the Options tab. Copy-paste the values of the Render Settings > Type Properties > Interactive > Shelf Tools > Options tab.

For the Script tab, we need some adjustments:

import loptoolutils, loputils

node = loptoolutils.genericTool(kwargs, '$HDA_NAME', '$HDA_NAME'.split("::")[1] + '_edit1')
node.parm('createprims').set('off')
loputils.setAllControlParameters(node, 'none')

As you can see, we need to split the name: $HDA_NAME will return fxquinox::contextinfo , and : is an unauthorized character for a Houdini node name. We simply isolate the second name component, and add the edit_1 part to it.

This is what you should end up with inside the Options tab:

And inside the Script tab:

You can now try to press TAB inside your Network View to drop the node, and should be welcomed by your HDA and its Edit variant:

Custom Icon

You can add an icon in the $HOUDINI_PATH following this naming convention: SCENEGRAPH_primtype_<custom schema class>.svg , to be displayed inside the Houdini Scene Graph Tree every time you apply your custom schema to a primitive:

In this case, the icon is SCENEGRAPH_primtype_fxquinoxcontextinfo.svg .