Skip to main content

Build nodes trees in Blender more elegantly with code

Project description

nodebpy

Run Tests

A package to help build node trees in blender more elegantly with python code.

The Design Idea

A text-based version of nodes should bring the convenience of writing code with IDE auto-completion, type hinting, with overall compactness and readability, while staying as close as possible to what building a node tree via the GUI feels like.

Other projects have attempted similar but none quite handled the API how I felt it should be done. Notable existing projects are:

Previous projects mostly implement the chaining of nodes together via class methods and chaining the . operator (instance_on_points().set_position()).

This approach is limiting - not being able to explicitly specify output sockets and input sockets while chaining.

nodebpy and the >> operator

In nodebpy we use the >> operator to link from one node or socket into another. This should feel and behave much like the Alt + Right Click drag between nodes in Node Wrangler. It will use some smart logic to match the most compatible sockets between the nodes, but if you ever want to be explicit you do so. The input and output sockets of a node are accessible as properties via the i_* and o_* prefixes, or you can use the ... placeholder to specify the particular input to be user, or pass in the previous node as a named argument.

g.Vector() >> g.SetPosition().i_offset
g.Vector() >> g.SetPosition(offset=...)
g.SetPosition(offset=g.Vector())

The >> operator will always look for the most compatible sockets first (matching data types) before looking for other compatible but not identical socket data types to link. If a compatible match can't be found an error will be thrown.

Contexts

What node tree or node tree interface we are currently editing is determined based on contexts. Instantiating a node class outside of a tree context will throw an error. The easiest way to enter and exit a tree context is to use the with statement.

Each time you instantiate a node class, a new node will be created and added to the current tree. If these nodes are given as arguments to other nodes or used with the >> operator, they will be automatically linked to the appropriate sockets.

Entering the tree.inputs and tree.outputs contexts will let you add new interface sockets to the node tree for usage as a modifier or as a node group in another node tree. These also return an object that can be used as arguments to other nodes or with the >> operator for linking.

with TreeBuilder("MyTree") as tree:
    points = g.Points(position=g.RandomValue.vector(min=-1))
    with tree.outputs:
        points >> s.SocketGeometry("New Points")

Example Node Tree

The node tree below creates a integer input and geometry output to the node group. We create a rotation variable that can be used later on as an argument, then construct a longer chain of nodes being created and linked together. The nodes are added and linked as each node is instantiated. After we exit the tree context, the nodes are automatically arranged.

from nodebpy import TreeBuilder, geometry as g, sockets as s

with TreeBuilder("AnotherTree", collapse=True) as tree:
    with tree.inputs:
        count = s.SocketInt("Count", 10)
    with tree.outputs:
        instances = s.SocketGeometry("Instances")

    rotation = (
        g.RandomValue.vector(min=-1, seed=2)
        >> g.AlignRotationToVector()
        >> g.RotateRotation(rotate_by=g.AxisAngleToRotation(angle=0.3))
    )

    _ = (
        count
        >> g.Points(position=g.RandomValue.vector(min=-1))
        >> g.InstanceOnPoints(instance=g.Cube(), rotation=rotation)
        >> g.SetPosition(
            position=g.Position() * 2.0 + (0, 0.2, 0.3),
            offset=(0, 0, 0.1),
        )
        >> g.RealizeInstances()
        >> g.InstanceOnPoints(g.Cube(), instance=...)
        >> instances
    )

Nodes

Documentation for all of the nodes can be found in the API Reference. This is mostly built automatically from the existing Blender node classes.

Every node has all of it's input sockets and enum options exposed as arguments to the class constructor. Input sockets are prefixed with i_ and output sockets are prefixed with o_. Properties that aren't exposed as sockets are available as class properties. Many properties are also available as class methods for convenience when constructing.

The basic math operators also automatically add relevant nodes with their operations and values set.

# operation is exposed as a property
math = g.Math(1.0, 2.0, operation='ADD')
math.operation = "SUBTRACT"

# operation can be chose as a class method
math = g.Math.subtract(1.0, 2.0)
math = g.Math.add(1.0, 2.0)

# these are equivalent, the g.Math.multiply is automatically added
g.Value(1.0) * 2
g.Math.multiply(g.Value(1.0), 2.0)

Design Considerations

Whenever possible, support IDE auto-complete and have useful types. We should know as much ahead of time as possible if our network will actually build.

  • Stick as closely to Geometry Nodes naming as possible
    • RandomValue creates a random value node
      • RandomValue.vector() creates it set to "VECTOR" data type and provides arguments for IDE auto-complete
  • Inputs and outputs from a node are prefixed with i_* and o_:
    • AccumulateField().o_total returns the output Total socket
    • AccumulateField().i_value returns the input Value socket
  • If inputs are subject to change depending on enums, provide separate constructor methods that provide related inputs as arguments. There should be no guessing involved and IDEs should provide documentation for what is required:
    • TransformGeometry.matrix(CombineTrasnsform(translation=(0, 0, 1))
    • TransformGeoemtry.components(translation=(0, 0, 1))
    • TransformGeometry(translation=(0, 0, 1))

Building

Most node classes are generated automatically with this. The nodes in nodes/manual.py are currently manually specified due to varying complexities of particular nodes (usually lergacy).

uv run generate.py && ruff format && ruff check --fix

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nodebpy-0.9.0.tar.gz (147.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

nodebpy-0.9.0-py3-none-any.whl (177.5 kB view details)

Uploaded Python 3

File details

Details for the file nodebpy-0.9.0.tar.gz.

File metadata

  • Download URL: nodebpy-0.9.0.tar.gz
  • Upload date:
  • Size: 147.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for nodebpy-0.9.0.tar.gz
Algorithm Hash digest
SHA256 97eb209504c499d20703b288a40dd03f792274c1187315e87b3d3e3d24342b19
MD5 05f8fe8c77d24cd3c422aa7fb674bea9
BLAKE2b-256 91d39944e6867164449adceec579e6448dfcb54ea5dab743e501a76e3f295b62

See more details on using hashes here.

Provenance

The following attestation bundles were made for nodebpy-0.9.0.tar.gz:

Publisher: pypi.yml on BradyAJohnston/nodebpy

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file nodebpy-0.9.0-py3-none-any.whl.

File metadata

  • Download URL: nodebpy-0.9.0-py3-none-any.whl
  • Upload date:
  • Size: 177.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for nodebpy-0.9.0-py3-none-any.whl
Algorithm Hash digest
SHA256 83c503704c48850f216aa3c16f508070d1fb962465ef580b91bbb4efa59b6d68
MD5 6cf95842e14785a14bc689614ef6900c
BLAKE2b-256 6c3ca64e9ddb65f715ccbfde15e4304fa404e7e7c014c31fe975e279d91fa3da

See more details on using hashes here.

Provenance

The following attestation bundles were made for nodebpy-0.9.0-py3-none-any.whl:

Publisher: pypi.yml on BradyAJohnston/nodebpy

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page