Python binding of TinyUSDZ

Currently in testing stage. Does not work well.

Core part is deletegated to native module(ctinyusd.so)

W.I.P.

Requirements

• Python 3.7 or later
  • Python 3.10+ recommended
  • For developping and testing, Python 3.8 or later is required.

Recommended python modules

• numpy
  • For efficient Attribute data handling.
  • from_numpy and to_numpy method available when numpy is installed..
• pandas
  • To support TimeSamples data efficiently(e.g. read/write to CSV, Excel)
• typeguard
  • To do type check at runtime.

Structure

• ctinyusdz: Native C++ module of tinyusdz
  • Python binding on top of C binding of TinyUSDZ.
• tinyusdz: Python module. Wraps some functions of ctinyusdz

TinyUSDZ's Python binding approach is like numpy: Frontend is written in Python for better Python integration(type hints for lsp(Intellisense), debuggers, exceptions, ...), and calls native C modules as necessary.

Supported platform

• [x] Linux
  • [x] x86-64
  • [x] aarch64
• [x] Windows
  • [x] 64bit AMD64
  • [x] 32bit x86
  • [ ] ARM64(not intensively tested, but should work)
• [x] macOS
  • [x] arm64
  • [x] Intel

Features

• T.B.W.

Optional

• [ ] pxrUSD compatible Python API?(pxr_compat folder)

Install through PyPI

$ python -m pip install tinyusdz

For developers. Build from source

Back to tinyusdz's root directory, then

# Use `build` module(install it with `python -m pip install build`) 
$ python -m build .

If you are working on TinyUSDZ Python module, Using setup.py recommended.

# install dependencies
$ python -m pip install setuptools scikit-build cmake ninja

$ python setup.py build
# Then copy `./_skbuild/-/cmake-install/tinyusdz/ctinyusdz.*.so/dll to `/python` folder.

Re-generate ctinyusdz.py

When TinyUSDZ C API has been updated, need to re-genrerate /python/tinyusdz/ctinyusd.py using ctypesgen https://github.com/ctypesgen/ctypesgen .

# if you do not install ctypesgen
$ python -m pip install ctypesgen


$ cd /python
$ sh gen-ctypes.sh 

Asan support

If you built ctinyusdz with ASAN enabled, use LD_PRELOAD to load asan modules.

LD_PRELOAD=/path/to/clang+llvm-14.0.0-x86_64-linux-gnu-ubuntu-18.04/lib/clang/14.0.0/lib/x86_64-unknown-linux-gnu/libclang_rt.asan.so  python tutorial.py

Please see https://tobywf.com/2021/02/python-ext-asan/ for more infos.

License

Apache 2.0 license.

EoL.