pycollada

0.9.3 · active · verified Sat Apr 11

pycollada is a Python library for reading, writing, and creating COLLADA (COLLAborative Design Activity) documents, an interchange file format for interactive 3D applications. It allows users to load a COLLADA file and interact with it as a Python object, supporting creation from scratch and in-place editing. The current version is 0.9.3, released on January 24, 2026, with a fairly active release cadence addressing compatibility and performance improvements.

Warnings

breaking Python 3.9 support was removed in pycollada 0.9.3, as Python 3.9 reached end-of-life. Python versions prior to 3.8 were dropped in 0.8.
Fix: Ensure you are using a supported Python version. For pycollada 0.9.3, Python 3.14+ is supported, and for 0.8+, Python 3.8+ (up to 3.12 for 0.8) is required. Python 3.9 is not supported by 0.9.3.
breaking NumPy became an explicit install requirement starting from version 0.8. Prior to this, it might have been an implicit dependency or optional. Additionally, versions 0.9.2 and 0.9 included fixes for compatibility with NumPy 2.0/2.3 after `fromstring` byte deprecation.
Fix: Always explicitly install `numpy` alongside `pycollada` (`pip install pycollada numpy`). If encountering issues with NumPy versions, ensure you have the latest `pycollada` (0.9.2 or newer for NumPy 2.x compatibility).
gotcha Loading large COLLADA XML documents might fail without explicit `huge_tree` support enabled for the underlying `lxml` parser. This was addressed in version 0.8.
Fix: Upgrade to pycollada 0.8 or newer. If you are unable to upgrade and dealing with large XML, you might need to manually configure `lxml` parsing options if pycollada exposes them, or consider pre-processing the DAE file.
gotcha When loading COLLADA documents, `pycollada` can raise exceptions like `DaeUnsupportedError` or `DaeBrokenRefError`. By default, these can halt loading.
Fix: To ignore specific errors and attempt to continue loading, pass a list of exception types to the `ignore` parameter of the `Collada` constructor: `mesh = Collada('file.dae', ignore=[DaeUnsupportedError, DaeBrokenRefError])`. Any errors encountered will be available in `Collada.errors`.
gotcha While pycollada 0.9.3 introduced compatibility with Python's built-in `ElementTree` API, `lxml` is still the recommended XML parser for better performance and robustness.
Fix: For optimal performance and to avoid potential parsing quirks, ensure `lxml` is installed (`pip install lxml`).

Install

pip install pycollada Install pycollada

Imports

Collada
```
from collada import Collada
```
The primary class for interacting with COLLADA documents is `Collada`, typically imported directly.
*
```
from collada import *
```
Common for quick access to all pycollada components in small scripts or interactive sessions.

Quickstart

This quickstart demonstrates how to create a simple COLLADA document containing a single red triangle and then save it. It also shows how to load an existing COLLADA file and access its basic properties. Requires numpy to be installed.

import os
from collada import Collada, material, geometry, source
import numpy as np

# 1. Create a new COLLADA document
mesh = Collada()

# 2. Define geometry: a simple triangle
# Vertices
verts = np.array([
    0.0, 0.0, 0.0,
    1.0, 0.0, 0.0,
    0.0, 1.0, 0.0
])
normals = np.array([
    0.0, 0.0, 1.0,
    0.0, 0.0, 1.0,
    0.0, 0.0, 1.0
])

vert_src = source.FloatSource('my_verts-positions', verts, ('X', 'Y', 'Z'))
norm_src = source.FloatSource('my_normals-normals', normals, ('X', 'Y', 'Z'))

# Indices for a single triangle
triangles = np.array([0, 0, 1, 1, 2, 2]) # vertex index, normal index, ...

geom = geometry.Geometry(mesh, 'geometry0', 'my_triangle_geometry',
                       [geometry.Triangles(vert_src, norm_src, triangles, 'material0')])

mesh.geometries.append(geom)

# 3. Create a material
effect = material.Effect('effect0', [], 'phong', diffuse=(1, 0, 0, 1))
mat = material.Material('material0', 'red_material', effect)
mesh.materials.append(mat)
mesh.effects.append(effect)

# 4. Save the COLLADA document
output_file = 'my_first_collada.dae'
mesh.write(output_file)
print(f"Created '{output_file}'")

# 5. Load an existing COLLADA document (or the one we just created)
# For demonstration, we'll try to load the file we just saved.
if os.path.exists(output_file):
    loaded_mesh = Collada(output_file)
    print(f"Loaded '{output_file}' with {len(loaded_mesh.geometries)} geometries.")
    # Example: Accessing a geometry
    if loaded_mesh.geometries:
        print(f"First geometry ID: {loaded_mesh.geometries[0].id}")
else:
    print(f"Could not load '{output_file}'. Make sure the file exists.")

view raw JSON →