Python library for reading, writing and managing 3D objects in the Khronos Group gltf and gltf2 formats.
pip install pygltflib==1.16.3
This is a library for reading, writing and handling GLTF files. It works for Python3.6 and above.
It supports the entire specification, including materials and animations. Main features are:
pip install pygltflib
from pygltflib import GLTF2
gltf = GLTF2()
from pygltflib import GLTF2, Scene
gltf = GLTF2()
scene = Scene()
gltf.scenes.append(scene) # scene available at gltf.scenes[0]
filename = "glTF-Sample-Models/2.0/AnimatedCube/glTF/AnimatedCube.gltf"
gltf = GLTF2().load(filename)
glb_filename = "glTF-Sample-Models/2.0/Box/glTF-Binary/Box.glb"
glb = GLTF2().load(glb_filename) # load method auto detects based on extension
glb = GLTF2().load_binary("BinaryGLTF.glk") # load_json and load_binary helper methods
gltf = GLTF2().load("glTF-Sample-Models/2.0/Box/glTF/Box.gltf")
current_scene = gltf.scenes[gltf.scene]
node_index = current_scene.nodes[0] # scene.nodes is the indices, not the objects
box = gltf.nodes[nodex_index]
box.matrix # will output vertices for the box object
from pygltflib import GLTF2, BufferFormat
gltf = GLTF2().load("glTF-Sample-Models/2.0/Box/glTF/Box.gltf")
gltf.convert_buffers(BufferFormat.BINARYBLOB) # convert buffers to GLB blob
gltf.convert_buffers(BufferFormat.DATAURI) # convert buffer URIs to data.
gltf.convert_buffers(BufferFormat.BINFILE) # convert buffers to files
gltf.save("test.gltf") # all the buffers are saved in 0.bin, 1.bin, 2.bin.
from pygltflib.utils import glb2gltf, gltf2glb
# convert glb to gltf
glb2gltf("glTF-Sample-Models/2.0/Box/glTF-Binary/Box.glb")
# on a primitve
gltf.meshes[0].primitives[0].extensions['KHR_draco_mesh_compression']
# on a material
gltf.materials[0].extensions['ADOBE_materials_thin_transparency']
# Application-specific semantics must start with an underscore, e.g., _TEMPERATURE.
a = Attributes()
a._MYCUSTOMATTRIBUTE = 123
gltf.meshes[0].primitives[0].attributes._MYOTHERATTRIBUTE = 456
gltf.remove_bufferView(0) # this will update all accessors, images and sparse accessors to remove the first bufferView
from pygltflib import GLTF2
from pygltflib.validator import validate, summary
filename = "glTF-Sample-Models/2.0/AnimatedCube/glTF/AnimatedCube.gltf"
gltf = GLTF2().load(filename)
validate(gltf) # will throw an error depending on the problem
summary(gltf) # will pretty print human readable summary of errors
# NOTE: Currently this experimental validator only validates a few rules about GLTF2 objects
from pygltflib import GLTF2
from pygltflib.utils import ImageFormat
filename = "glTF-Sample-Models/2.0/AnimatedCube/glTF/AnimatedCube.gltf"
gltf = GLTF2().load(filename)
gltf.convert_images(ImageFormat.FILE)
gltf.images[0].uri # will now be 0.png and the texture image will be saved in 0.png
from pygltflib import GLTF2
from pygltflib.utils import ImageFormat
filename = "glTF-Sample-Models/2.0/AnimatedCube/glTF/AnimatedCube.gltf"
gltf = GLTF2().load(filename)
gltf.images[0].name = "cube.png" # will save the data uri to this file (regardless of data format)
gltf.convert_images(ImageFormat.FILE)
gltf.images[0].uri # will now be cube.png and the texture image will be saved in cube.png
from pygltflib import GLTF2
from pygltflib.utils import ImageFormat, Image
gltf = GLTF2()
image = Image()
image.uri = "myfile.png"
gltf.images.append(image)
gltf.convert_images(ImageFormat.DATAURI)
gltf.images[0].uri # will now be something like "data:image/png;base64,iVBORw0KGg..."
gltf.images[0].name # will be myfile.png
This is an unofficial library that tracks the official file format for GLTF2.
The library was initially built to load and save simple meshes but support for the entire spec, including materials and animations is pretty good. Supports both json (.gltf) and binary (.glb) file formats, although .glb support is missing some features at the moment.
It requires python 3.6 and above because it uses dataclasses and all attributes are type hinted. And f-strings, plenty of f-strings.
Check the table below for an idea of which sample models validate.
Questions? Contributions? Bug reports? Open an issue on the gitlab page for the project.
We are very interested in hearing your use cases for pygltflib
to help drive the roadmap.
pygltflib
made for 'The Beat: A Glam Noir Game' supported by Film Victoria.
1.13.5
utils.validator
by moving to validator.validate
validator.summary
to provide human readable output of validator1.13.4
remove_bufferView
if leaving dangling references to removed bufferViewremove_bufferView
metallicRoughnessTexture
to use TextureInfo
remove_bufferView
when sparse accessor is NoneInvalidAccessorSparseIndicesComponentTypeException
1.13.3
GLTF.convert_images
to convert from buffers to image files. NOTE: Does not update buffer yet.GLTF.convert_images
to convert from buffers to data uris. NOTE: Does not update buffer yet.1.13.2
1.13.1
GLTF.convert_images
method for converting texture image data from data uris to files and vice versaImage
Primitive
so that Attributes
is created on init1.13.0
SparseAccessor
in favour of AccessorSparseIndices
and AccessorSparseValues
(please update your code)MaterialTexture
in favour of TextureInfo
to better match GLTF2 specification (please update your code)AlphaMode
ENUM in favour of constants (eg replace AlphaMode.OPAQUE
with OPAQUE
(please update your code)material.occlusionTextureInfo
material.normalTextureInfo
Animation
class by adding AnimationSampler
utils.validator
to validate GLTF2 object accessor and bufferViews1.12.0
1.11.10
1.11.9
1.11.8
1.11.7
normalized
flag to Accessor1.11.6
None
by default (use primitive.attributes = Attributes()
)1.11.5
identify_uri
to avoid using lengthy byte strings as file names1.11.4
1.11.3
1.11.2
1.11.1
1.11
GLTF.binary_blob()
convert_buffers
method to switch buffer formats between data uri, binary files and binary blobs1.10
1.9
1.8
1.7
1.6
load_json
, load_binary
, save_json
and save_binary
from internal methods1.5
glb2gltf
and gltf2glb
util functions to pygltflib.utils
for easy file conversion1.4
pygltfib.utils
1.3
1.2
1.0
pip install pygltflib
git clone https://gitlab.com/dodgyville/pygltflib
Note: These examples use the official sample models provided by Khronos at:
https://github.com/KhronosGroup/glTF-Sample-Models
from pygltflib import *
# create gltf objects for a scene with a primitive triangle with indexed geometry
gltf = GLTF2()
scene = Scene()
mesh = Mesh()
primitive = Primitive()
node = Node()
buffer = Buffer()
bufferView1 = BufferView()
bufferView2 = BufferView()
accessor1 = Accessor()
accessor2 = Accessor()
# add data
buffer.uri = "data:application/octet-stream;base64,AAABAAIAAAAAAAAAAAAAAAAAAAAAAIA/AAAAAAAAAAAAAAAAAACAPwAAAAA="
buffer.byteLength = 44
bufferView1.buffer = 0
bufferView1.byteOffset = 0
bufferView1.byteLength = 6
bufferView1.target = ELEMENT_ARRAY_BUFFER
bufferView2.buffer = 0
bufferView2.byteOffset = 8
bufferView2.byteLength = 36
bufferView2.target = ARRAY_BUFFER
accessor1.bufferView = 0
accessor1.byteOffset = 0
accessor1.componentType = UNSIGNED_SHORT
accessor1.count = 3
accessor1.type = SCALAR
accessor1.max = [2]
accessor1.min = [0]
accessor2.bufferView = 1
accessor2.byteOffset = 0
accessor2.componentType = FLOAT
accessor2.count = 3
accessor2.type = VEC3
accessor2.max = [1.0, 1.0, 0.0]
accessor2.min = [0.0, 0.0, 0.0]
primitive.attributes.POSITION = 1
node.mesh = 0
scene.nodes = [0]
# assemble into a gltf structure
gltf.scenes.append(scene)
gltf.meshes.append(mesh)
gltf.meshes[0].primitives.append(primitive)
gltf.nodes.append(node)
gltf.buffers.append(buffer)
gltf.bufferViews.append(bufferView1)
gltf.bufferViews.append(bufferView2)
gltf.accessors.append(accessor1)
gltf.accessors.append(accessor2)
# save to file
gltf.save("triangle.gltf")
pygltflib
can load json-based .GLTF files and binary .GLB files, based on the file extension.
>>> from pygltflib import GLTF2
>>> filename = "glTF-Sample-Models/2.0/AnimatedCube/glTF/AnimatedCube.gltf"
>>> gltf = GLTF2().load(filename)
>>> gltf.scene
0
>>> gltf.scenes
[Scene(name='', nodes=[0])]
>>> gltf.nodes[0]
Node(mesh=0, skin=None, rotation=[0.0, -1.0, 0.0, 0.0], translation=[], scale=[], children=[], matrix=[], camera=None, name='AnimatedCube')
>>> gltf.nodes[0].name
'AnimatedCube'
>>> gltf.meshes[0].primitives[0].attributes
Attributes(NORMAL=4, POSITION=None, TANGENT=5, TEXCOORD_0=6)
>>> filename2 = "test.gltf"
>>> gltf.save(filename2)
>>> from pygltflib import GLTF2
>>> glb_filename = "glTF-Sample-Models/2.0/Box/glTF-Binary/Box.glb"
>>> glb = GLTF2().load(glb_filename)
>>> glb.scene
0
>>> glb.scenes
[Scene(name='', nodes=[0])]
>>> glb.nodes[0]
Node(mesh=None, skin=None, rotation=[], translation=[], scale=[], children=[1], matrix=[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, -1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0], camera=None, name=None)
>>> glb.meshes[0].primitives[0].attributes
Attributes(POSITION=2, NORMAL=1, TANGENT=None, TEXCOORD_0=None, TEXCOORD_1=None, COLOR_0=None, JOINTS_0=None, WEIGHTS_0=None)
>>> glb.save("test.glb")
>>> glb.binary_blob() # read the binary blob used by the buffer in a glb
<a bunch of binary data>
>>> from pygltflib import GLTF2
>>> # convert glb to gltf
>>> glb = GLTF2().load("glTF-Sample-Models/2.0/Box/glTF-Binary/Box.glb")
>>> glb.save("test.gltf")
>>> # convert gltf to glb
>>> gltf = GLTF2().load("glTF-Sample-Models/2.0/Box/glTF/Box.gltf")
>>> gltf.save("test.glb")
>>> from pygltflib import GLTF2
>>> from pygltflib.utils import glb2gltf, gltf2glb
>>> # convert glb to gltf
>>> glb2gltf("glTF-Sample-Models/2.0/Box/glTF-Binary/Box.glb")
>>> # convert gltf to glb
>>> gltf2glb("glTF-Sample-Models/2.0/Box/glTF/Box.gltf", "test.glb", override=True)
The data for a buffer in a GLTF2 files can be stored in the buffer object's URI string or in a binary file pointed to by the buffer objects' URI string or as a binary blob inside a GLB file.
While saving and loading GLTF2 files is mostly handled transparently by the library, there may be some situations where you want a specific type of buffer storage.
For example, if you have a GLTF file that stores all the associated data in .bin files but you want to create a single file, you need to convert the buffers from binary files to data uris or glb binary data.
There is a convenience method named convert_buffers
that can help.
>>> from pygltflib import GLTF2, BufferFormat
>>> gltf = GLTF2().load("glTF-Sample-Models/2.0/Box/glTF/Box.gltf")
>>> gltf.convert_buffers(BufferFormat.DATAURI) # convert buffer URIs to data.
>>> gltf.save_binary("test.glb") # try and save, will get warning.
Warning: Unable to save data uri to glb format.
>>> gltf.convert_buffers(BufferFormat.BINARYBLOB) # convert buffers to GLB blob
>>> gltf.save_binary("test.glb")
>>> gltf.convert_buffers(BufferFormat.BINFILE) # convert buffers to files
>>> gltf.save("test.gltf") # all the buffers are saved in 0.bin, 1.bin, 2.bin.
The image data for textures in GLTF2 files can be stored in the image objects URI string or in an image file pointed to by the image objects' URI string or as part of the buffer.
While saving and loading GLTF2 files is mostly handled transparently by the library, there may be some situations where you want a specific type of image storage.
For example, if you have a GLB file that stores all its image files in .PNG files but you want to create a single GLTF file, you need to convert the images from files to data uris.
Currently converting images to and from the buffer is not supported. Only image files and data uris are supported.
There is a convenience method named convert_images
that can help.
>>> # embed an image file to your GLTF.
>>> from pygltflib.utils import ImageFormat, Image
>>> gltf = GLTF2()
>>> image = Image()
>>> image.uri = "myfile.png"
>>> gltf.images.append(image)
>>> gltf.convert_images(ImageFormat.DATAURI) # image file will be imported into the GLTF
>>> gltf.images[0].uri # will now be something like "data:image/png;base64,iVBORw0KGg..."
>>> gltf.images[0].name # will be myfile.png
>>> # create an image file from GLTF data uris
>>> from pathlib import Path
>>> from pygltflib.utils import ImageFormat, Image
>>> gltf = GLTF2()
>>> image = Image()
>>> image.uri = "data:image/png;base64,iVBORw0KGg..."
>>> image.name = "myfile.png" # optional file name, if not provided, the image files will be called "0.png", "1.png"
>>> gltf.images.append(image)
>>> gltf.convert_images(ImageFormat.FILE) # image file will be imported into the GLTF
>>> gltf.images[0].uri # will be myfile.png
"myfile.png"
>>> Path("myfile.png").exists()
True
The GLTF2 spec allows for extensions to added to any component of a GLTF file.
As of writing (August 2019) there are about a dozen extensions from Khronos and other vendors
In pygltflib, extensions are loaded as ordinary dict
objects and so should be accessed like regular key,value pairs.
For example extensions["KHR_draco_mesh_compression"]["bufferView"]
instead of extensions["KHR_draco_mesh_compression"].bufferView
.
This allows future extensions to be automatically supported by pygltflib.
Extras should work the same way.
Using sample models loaded and then saved using this library, here are validator reports (blank is untested). If available, The result of a visual inspection is in brackets next to the validator result.
Model | gltf to gltf | gltf to glb | glb to gltf | glb to glb |
---|---|---|---|---|
2CylinderEngine | passes | passes | passes | passes |
AlphaBlendModeTest | passes | passes | passes | passes |
AnimatedCube | passes | passes | no glb available | no glb available |
AnimatedMorphCube | passes | passes | passes | passes |
AnimatedMorphSphere | passes | passes | passes | passes |
AnimatedTriangle | passes | passes | no glb available | no glb available |
Avocado | passes | passes | passes | passes |
BarramundiFish | passes | passes | passes | passes |
BoomBox | passes | passes | passes | passes |
BoomBoxWithAxes | passes | passes | no glb available | no glb available |
Box | passes | passes | passes | passes |
BoxAnimated | passes | passes | passes | |
BoxInterleaved | passes | passes | passes | |
BoxTextured | passes | passes | ||
BoxTexturedNonPowerOfTwo | passes | passes | ||
BoxVertexColors | passes | passes | ||
BrainStem | passes | passes | passes | |
Buggy | passes | passes | passes | |
Cameras | passes | passes | no glb available | no glb available |
CesiumMan | passes | passes | ||
CesiumMilkTruck | passes | passes | ||
Corset | passes | passes | passes | passes |
Cube | passes | passes | no glb available | no glb available |
DamagedHelmet | passes | passes | passes | passes |
Duck | passes | passes | passes | passes |
FlightHelmet | passes | passes | no glb available | no glb available |
GearboxAssy | passes | passes | ||
Lantern | passes | passes | ||
MetalRoughSpheres | passes | passes | ||
Monster | passes | passes | ||
MultiUVTest | passes | passes | ||
NormalTangentMirrorTest | passes | passes | ||
NormalTangentTest | passes | passes | passes | |
OrientationTest | passes | passes | ||
ReciprocatingSaw | passes | passes | ||
RiggedFigure | passes | passes | ||
RiggedSimple | passes | passes | ||
SciFiHelmet | passes | passes | no glb available | no glb available |
SimpleMeshes | passes | passes | no glb available | no glb available |
SimpleMorph | passes | passes | no glb available | no glb available |
SimpleSparseAccessor | passes | passes | no glb available | no glb available |
SpecGlossVsMetalRough | passes | passes | passes | passes |
Sponza | passes | passes | no glb available | no glb available |
Suzanne | passes | passes | no glb available | no glb available |
TextureCoordinateTest | passes | passes | passes | passes |
TextureSettingsTest | passes | passes | passes | passes |
TextureTransformTest | passes | passes | no glb available | no glb available |
Triangle | passes | passes | no glb available | no glb available |
TriangleWithoutIndices | passes | passes | no glb available | no glb available |
TwoSidedPlane | passes | passes | no glb available | no glb available |
VC | passes | fails | passes | passes |
VertexColorTest | passes | passes | passes | passes |
WaterBottle | passes | passes | passes | passes |
What does pygltflib.utils.validator test? NOTE: At the moment the validator raises an exception when an rule is broken. If you have ideas of the best way to return information on validation warnings/errors please open a ticket on our gitlab.
Rule | validator tests | exception raised |
---|---|---|
accessor.componentType must be valid | yes | InvalidAcccessorComponentTypeException |
accessor min and max arrays must be valid length | yes | InvalidArrayLengthException |
accessor min and max arrays must be same length | yes | MismatchedArrayLengthException |
mesh.primitive.mode must be valid | yes | InvalidMeshPrimitiveMode |
accessor.sparse.indices.componentType must be valid | yes | InvalidAccessorSparseIndicesComponentTypeException |
bufferView byteOffset and byteStrides must be valid | yes | InvalidValueError |
bufferView targets must be valid | yes | InvalidBufferViewTarget |
all other tests | no |
git clone https://github.com/KhronosGroup/glTF-Sample-Models
pytest test_pygltflib.py