You can think of Named Graphs as Bifrost Compounds for Maya Native nodes. Where you create a Node Graph in Maya as Save it out as a reusable node graph in Json Format. These Graphs can be loaded back in any time and edited then saved anytime(See Example).


Lets first have a look at a simple maya graph we’re going save to json

alternate text

We have 3 nodes in the center which we want to save.

First select them

Now run the following to convert them to a simple dict which provides the ID and the node instance.


While here we’re simply using the node name for the id, you can however use any id you would like.

from zoo.libs.maya import zapi
from zoo.libs.hive.base.serialization import dggraph
nodes = { i for i in zapi.selected()}

Now lets create some inputs and outputs for our graph

inputs = {"startMtx": nodes["midRotateOnly"].inputMatrix,
          "endMtx": nodes["endTranslateScaleOnly"].inputMatrix}
outputs = {"outputMatrix": nodes["tangentEndWorldMtx"].matrixSum}

Now lets create a graph instance with our scene data and serialize to a dict read to save

graph = dggraph.NamedDGGraph(graphId="ourGraphId", name="ourGraphName",
                             metaData, inputs, outputs)


“ourGraphId” is quite important as this is used as the registered Id for later retrieval

Now save to the registry

from zoo.libs.hive import api

outputPath = api.Configuration().graphRegistry().saveGraph("ourGraphId", graph)

Alright we’ve seen how to save a maya graph out, now let’s recreate it and have a look at some nifty options.

from zoo.libs.hive import api
dataGraphRep = api.Configuration().graphRegistry().graph("ourGraphId")
sceneGraph = api.serialization.NamedDGGraph.create(dataGraphRep)
alternate text

We’ve now created the graph but don’t you think it would be nice to have inputs/outputs of the graph as specialized nodes for easier troubleshooting and editing? Well lets do that.

alternate text

As you can see we now have a input node(left) and a output node(right). This is great for editing our graph and troubleshooting our connections.

Now lets add a new input just for our demo, also you can use addOutput method as well.

sceneGraph.addInput("myNewInput", sceneGraph.node("midRotateOnly").useTranslate)
sceneGraph.addInput("myNewInput", sceneGraph.node("endTranslateScaleOnly").useTranslate)
alternate text

Here we add one inputAttribute by connected to 2 internal attributes using the helper method sceneGraph.node which retrieves a node by it’s ID.

At this point we can re-save the graph

outputPath = api.Configuration().graphRegistry().saveGraph("ourGraphId", graph)

Now lets delete the IO nodes so we don’t incur any runtime performance hits. But lets first connect the output to a transform so you can see that you can create/edit and clean the graph at any time.

alternate text

Now delete.

alternate text