This PR does the following things

• Performed typo refactoring in pyx files
• Updated to newer Embree API (2.) . Embree 3.0 is being developed...
• Added the possibility to export all embree results when performing request
• Added 12 new tests run from nosetests, activated them in travis
• Run examples in travis

One can discuss each point...

Hi,

Thanks for making this git. Could you give some more details on how to install Pyembree?

In Ubuntu command line, I insert sudo python setup.py install

But there is some missing folder embree2 appartently... Or do I first have to install and compile embree itself?

Best regards, Arne

I have updated a trianges.pyx since it is using a missing attribute.

I guess one wants RTC_GEOMETRY_STATIC instead of RTCGEOMETRY_STATIC.

https://github.com/embree/embree/blob/90e49f243703877c7714814d6eaa5aa3422a5839/include/embree2/rtcore_geometry.h#L72

The original error log is presented here

D:\Embree\pyembree>python setup.py build
Please put "# distutils: language=c++" in your .pyx or .pxd file(s)
Compiling pyembree\trianges.pyx because it changed.
[1/1] Cythonizing pyembree\trianges.pyx

Error compiling Cython file:
------------------------------------------------------------
...
def run_triangles():
    pass

cdef unsigned int addCube(rtcs.RTCScene scene_i):
    cdef unsigned int mesh = rtcg.rtcNewTriangleMesh(scene_i,
                rtcg.RTCGEOMETRY_STATIC, 12, 8, 1)
                   ^
------------------------------------------------------------

pyembree\trianges.pyx:19:20: cimported module has no attribute 'RTCGEOMETRY_STATIC'
Traceback (most recent call last):
  File "setup.py", line 11, in <module>
    include_path=include_path)
  File "C:\Program Files\Python36\lib\site-packages\Cython\Build\Dependencies.py", line 1039, in cythonize
    cythonize_one(*args)
  File "C:\Program Files\Python36\lib\site-packages\Cython\Build\Dependencies.py", line 1161, in cythonize_one
    raise CompileError(None, pyx_file)
Cython.Compiler.Errors.CompileError: pyembree\trianges.pyx

I'd like to run Pyembree in an AWS Lambda function (via a Lambda 'Layer'), which means Embree will be located in /opt/python/embree. I'm having a bit of trouble configuring Pyembree to expect Embree in this location.

This is what I've tried so far (cobbled together from this script and this comment) to build the environment:

sudo amazon-linux-extras install python3.8
sudo yum install python38-devel gcc gcc-c++
wget https://github.com/embree/embree/releases/download/v2.17.7/embree-2.17.7.x86_64.linux.tar.gz -O /tmp/embree.tar.gz -nv
sudo mkdir /opt/python/embree
sudo tar -xzf /tmp/embree.tar.gz --strip-components=1 -C /opt/python/embree
sudo pip3.8 install --no-cache-dir numpy cython
wget https://github.com/scopatz/pyembree/releases/download/0.1.6/pyembree-0.1.6.tar.gz
tar xf pyembree-0.1.6.tar.gz
sed -i -e 's/embree2/\/opt\/python\/embree\/include\/embree2/g' pyembree-0.1.6/pyembree/*
tar czf pyembree-0.1.6.tar.gz pyembree-0.1.6
sudo pip3.8 install --global-option=build_ext --global-option="-I/opt/python/embree/include" --global-option="-L/opt/python/embree/lib" --target=/opt/python pyembree-0.1.6.tar.gz

This seems to build without problem and puts Embree and Pyembree in /opt/python. If I cd into /opt/python and run Python, I can import Pyembree, but the build can't find libembree.so.2:

Python 3.8.5 (default, Feb 18 2021, 01:24:20)
[GCC 7.3.1 20180712 (Red Hat 7.3.1-12)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import pyembree
>>> from pyembree import rtcore_scene as rtcs
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: libembree.so.2: cannot open shared object file: No such file or directory

Any idea what else I should try? I'm not sure if I should be replacing embree2 with opt/python/embree/include/embree2 before building the pxd/pyx files, for example. I've also tried altering setup.py to: include_path = [np.get_include(), "/opt/python/embree/include", "/opt/python/embree/lib"].

Any pointers very welcome!

Thanks for the great package! In a trimesh issue someone posted a backtrace that looked like it was occurring in the pyembree destructor, I was wondering if you'd ever seen anything similar?

Thread 1 "python" received signal SIGSEGV, Segmentation fault.
0x0000000000000000 in ?? ()
(gdb) py-bt
Traceback (most recent call first):
(gdb) bt
#0  0x0000000000000000 in ?? ()
#1  0x00007fffd8ab7c30 in embree::avx::TriangleMeshISA::~TriangleMeshISA() ()
   from /usr/local/lib/libembree.so.2
#2  0x00007fffd850002f in embree::Scene::~Scene() ()
   from /usr/local/lib/libembree.so.2
#3  0x00007fffd8500179 in embree::Scene::~Scene() ()
   from /usr/local/lib/libembree.so.2
#4  0x00007fffd84c3cc5 in rtcDeleteScene () from /usr/local/lib/libembree.so.2
#5  0x00007fffd992474c in __pyx_pf_8pyembree_12rtcore_scene_11EmbreeScene_4__dealloc__ (__pyx_v_self=0x7fffd3166490) at pyembree/rtcore_scene.cpp:3434
#6  __pyx_pw_8pyembree_12rtcore_scene_11EmbreeScene_5__dealloc__ (
    __pyx_v_self=<pyembree.rtcore_scene.EmbreeScene at remote 0x7fffd3166490>)
    at pyembree/rtcore_scene.cpp:3419
#7  __pyx_tp_dealloc_8pyembree_12rtcore_scene_EmbreeScene (
    o=<pyembree.rtcore_scene.EmbreeScene at remote 0x7fffd3166490>)
    at pyembree/rtcore_scene.cpp:6042
#8  0x00000000004fc70f in PyDict_Clear () at ../Objects/dictobject.c:946
#9  0x00000000005419b9 in dict_tp_clear.lto_priv.332 (op=<optimized out>)
    at ../Objects/dictobject.c:2152
#10 0x000000000049ca0f in delete_garbage (
    old=0x8fa280 <generations.lto_priv+96>, collectable=0x7fffffffdb40)
    at ../Modules/gcmodule.c:820
#11 collect.lto_priv () at ../Modules/gcmodule.c:984
---Type <return> to continue, or q <return> to quit---
#12 0x00000000004f9ade in PyGC_Collect () at ../Modules/gcmodule.c:1440
#13 0x00000000004f8d7f in Py_Finalize () at ../Python/pythonrun.c:448
#14 0x00000000004936f2 in Py_Main () at ../Modules/main.c:665
#15 0x00007ffff7810830 in __libc_start_main (main=0x4932b0 <main>, argc=2, 
    argv=0x7fffffffddd8, init=<optimized out>, fini=<optimized out>, 
    rtld_fini=<optimized out>, stack_end=0x7fffffffddc8)
    at ../csu/libc-start.c:291
#16 0x00000000004931d9 in _start ()

Using the output dict to get the distance to the intersection is very slow. So I added a new query type, distance, which returns just the distance to the hit.

Hi, thanks for the great library!

Someone opened an issue on trimesh about the errors that get printed when you allocate multiple scenes. It's not really a functional problem as pyembree still returns the correct result, I was wondering if there was a procedure or destructor I could call to suppress these warnings?

import numpy as np

from pyembree import rtcore_scene
from pyembree.mesh_construction import TriangleMesh

if __name__ == '__main__':
     triangles_a = np.random.random((10,3,3))
     scene_a = rtcore_scene.EmbreeScene()
     mesh_a = TriangleMesh(scene_a, triangles_a)

     # do something to deallocate here?

     triangles_b = np.random.random((10,3,3))
     scene_b = rtcore_scene.EmbreeScene()
     mesh_b = TriangleMesh(scene_b, triangles_b)

produces this warning:

ERROR CAUGHT IN EMBREE
ERROR: Invalid operation
ERROR MESSAGE: b'/home/benthin/Projects/embree_v251/kernels/common/rtcore.cpp (157): already initialized'

Best, Mike

in the same way as RTCFilterFunc in rtcore_geometry.pyx. This allows me to set custom intersection functions from cython code, in the same way that you already can with filter feedback functions:

    from mesh_intersection cimport patchIntersectFunc
    cimport pyembree.rtcore_geometry_user as rtcgu
    .
    .
    .
    rtcgu.rtcSetIntersectFunction(scene, geomID, <rtcgu.RTCIntersectFunc> patchIntersectFunc)

This pull request adds support for creating hexahedral and tetrahedral meshes. It also implements creating triangular meshes using an indices array as well as a vertices array.

Since Embree 3.13.0 (https://github.com/embree/embree/releases/tag/v3.13.0) Apple Silicon is supported with Embree. pyembree should be updated to support it. Also see: https://github.com/scopatz/pyembree/issues/28

Hi, when attempting to install pyembree in a Python3.9 environment I get an error due to incompatible packages (see code below). This was tested on a MacBook Pro (2017) running macOS 10.14.6. Is there any way to resolve this?

$ conda create --name python3.9 -c conda-forge python=3.9 pyembree
Collecting package metadata (current_repodata.json): done
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: |
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
failed

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions

Package python conflicts for:
python=3.9
pyembree -> numpy[version='>=1.18.1,<2.0a0'] -> python[version='3.7.*|3.8.*|>=3.9,<3.10.0a0']
pyembree -> python[version='2.7.*|3.5.*|3.6.*|>=2.7,<2.8.0a0|>=3.6,<3.7.0a0|>=3.8,<3.9.0a0|>=3.7,<3.8.0a0|>=3.5,<3.6.0a0|3.4.*']

https://github.com/scopatz/pyembree/blob/master/pyembree/mesh_construction.pyx#L158 This link seems to be dead. I suppose that the node ordering is something like [[0, 0, 0], [1, 0, 0], [1, 1, 0], [0, 1, 0], [0, 0, 1], [1, 0, 1], [1, 1, 1], [0, 1, 1]] for a unit cube, right?

[edit] same here: https://github.com/scopatz/pyembree/blob/master/pyembree/mesh_construction.h#L4