databpy

A set of data-oriented wrappers around the python API of Blender.

This was originally used internally inside of Molecular Nodes but was broken out into a separate python module for re-use in other projects.

Installation

Available on PyPI, install with pip:

pip install databpy

[!CAUTION]

bpy (Blender as a python module) is listed as an optional dependency, so that if you install databpy inside of Blender you won’t install a redundant version of bpy. If you are using this outside of Blender, you will need to specifically request bpy with either of these methods:

# install wtih bpy dependency
pip install "databpy[bpy]"

# install both packages
pip install databpy bpy

# install with all optional dependencies
pip install "databpy[all]"

Usage

The main use cases are to create objects, store and retrieve attributes from them. The functions are named around nodes in Geometry Nodes Store Named Attribute and Named Attribute

import databpy as db

# store and retrieve attributes from a mesh, point cloud, or curves object
db.store_named_attribute()
db.named_attribute()

Here’s an example on how to store an attribute:

import numpy as np
import databpy as db

coords = np.array([[0, 0, 0], [0, 5, 0], [5, 0, 0], [5, 5, 0]]))

obj = db.create_object(coords, name="Box")
db.store_named_attribute(obj, np.array([10, 20, 31, 42]), "vals")

This module is mainly used to create mesh objects and work with their attributes. It is built to store and retrieve data using NumPy arrays:

import numpy as np
import databpy as db

np.random.seed(6)

# Create a mesh object
random_verts = np.random.rand(10, 3)

obj = db.create_object(random_verts, name="RandomMesh")

obj.name

'RandomMesh'

Access attributes from the object’s mesh.

db.named_attribute(obj, "position")

array([[0.8928602 , 0.3319798 , 0.8212291 ],
       [0.04169663, 0.10765668, 0.59505206],
       [0.52981734, 0.41880742, 0.33540785],
       [0.62251943, 0.43814144, 0.7358821 ],
       [0.5180364 , 0.5788586 , 0.6453551 ],
       [0.99022424, 0.8198582 , 0.41320094],
       [0.8762677 , 0.82375944, 0.05447451],
       [0.7186372 , 0.8021706 , 0.7364066 ],
       [0.7091318 , 0.5409368 , 0.12482417],
       [0.9576473 , 0.4032563 , 0.21695116]], dtype=float32)

BlenderObject class (bob)

This is a convenience class that wraps around the bpy.types.Object, and provides access to all of the useful functions. We can wrap an existing Object or return one when creating a new object.

This just gives us access to the named_attribute() and store_named_attribute() functions on the object class, but also provides a more intuitive way to access the object’s attributes.

# wrap an existing object or create a new one
bob = db.BlenderObject(obj)
bob = db.create_bob(random_verts)

# these two are identical
bob.named_attribute("position")
bob.position

AttributeArray(name='position', object='NewObject', mesh='NewObject', domain=POINT, type=FLOAT_VECTOR, shape=(10, 3), dtype=float32)
array([[0.8928602 , 0.3319798 , 0.8212291 ],
       [0.04169663, 0.10765668, 0.59505206],
       [0.52981734, 0.41880742, 0.33540785],
       [0.62251943, 0.43814144, 0.7358821 ],
       [0.5180364 , 0.5788586 , 0.6453551 ],
       [0.99022424, 0.8198582 , 0.41320094],
       [0.8762677 , 0.82375944, 0.05447451],
       [0.7186372 , 0.8021706 , 0.7364066 ],
       [0.7091318 , 0.5409368 , 0.12482417],
       [0.9576473 , 0.4032563 , 0.21695116]], dtype=float32)

We can clear all of the data from the object and initialise a new mesh underneath:

bob.new_from_pydata(np.random.randn(5, 3))
bob.position

AttributeArray(name='position', object='NewObject', mesh='NewObject', domain=POINT, type=FLOAT_VECTOR, shape=(5, 3), dtype=float32)
array([[ 0.82465386, -1.1764315 ,  1.5644896 ],
       [ 0.7127051 , -0.1810066 ,  0.53419954],
       [-0.58661294, -1.4818532 ,  0.8572476 ],
       [ 0.94309896,  0.11444143, -0.02195668],
       [-2.1271446 , -0.83440745, -0.4655083 ]], dtype=float32)

Example with Polars data

import polars as pl
import databpy as db
from io import StringIO

json_file = StringIO("""
{
  "Dino": [
    [55.3846, 97.1795, 0.0],
    [51.5385, 96.0256, 0.0]
  ],
  "Star": [
    [58.2136, 91.8819, 0.0],
    [58.1961, 92.215, 0.0]
  ]
}
""")

df = pl.read_json(json_file)
columns_to_explode = [col for col in df.columns if df[col].dtype == pl.List(pl.List)]
df = df.explode(columns_to_explode)

vertices = np.zeros((len(df), 3), dtype=np.float32)
bob = db.create_bob(vertices, name="DinoStar")

for col in df.columns:
    data = np.vstack(df.get_column(col).to_numpy())
    bob.store_named_attribute(data, col)

bob.named_attribute("Dino")

array([[55.3846, 97.1795,  0.    ],
       [51.5385, 96.0256,  0.    ]], dtype=float32)

bob.named_attribute("Star")

array([[58.2136, 91.8819,  0.    ],
       [58.1961, 92.215 ,  0.    ]], dtype=float32)