如何在PyQt5中正确使用QOpenGLBuffer的分配/读取/写入？

Question

This需要一个void*数据输入。所以，我让它和这样的东西一起工作

vbo = QtGui.QOpenGLBuffer(QtGui.QOpenGLBuffer.VertexBuffer)
vrtxAttrs = [1., 1., 1., 0., 0.1, 0.1, 1.0,
             -1., -1., -1., 1.0, 0.9, 0.9, 1.0] # a list of interleaved positions xyz and colors rgba
data = numpy.array(vrtxAttrs, dtype=numpy.float32)

老实说，我不知道我是怎么来的。我尝试了很多东西，却无意中使用了data.ctypes.data_as(ctypes.POINTER(ctypes.c_void_p)).contents。请告诉我这很糟糕，还有更好的办法

# How I allocate..
vbo.allocate(data.ctypes.data_as(ctypes.POINTER(ctypes.c_void_p)).contents,
                          sys.getsizeof(data))

# How I write.. (I believe this calls glBufferSubData internally)
# I intend to change the second vertex's position and color.
newdata = [2., 2., 2., 0.5, 0.4, 0.4, 1.]
toWrite = numpy.array(newdata, dtype=numpy.float32)
offset = 1 * 7 * 4 # this has to be in bytes. (1 is the index, 7 the stride, 4 the bytesize.)
count = len(newdata) * 4 # this has to be in bytes.
vbo.write(offset, data.ctypes.data_as(ctypes.POINTER(ctypes.c_void_p)).contents, count)

# How I read,
# I don't read, since I don't know what weird cast I need to do to read.
# The usage is documented as bool QOpenGLBuffer::read(int offset, void *data, int count)

这些Python文档是荒谬的，是的，我们知道如何在C++中使用它们，Python呢？首先，我在寻找解释sip.voidptr的答案

Answer 1

这个答案可能不会赢得任何奖品，它是在凌晨4点匆忙完成的。😣

我无法从代码中确切了解您想要做什么，但据我所知，问题是如何在QOpenGLBuffer的上下文中使用allocate/read/write来处理NumPy数组

# Your imports / initializer code here...

buffer = QOpenGLBuffer(QOpenGLBuffer.VertexBuffer)
buffer.create() #Creates the buffer object in the OpenGL server
buffer.bind()   # Binds object to the current OpenGL context
buffer.allocate(120) # How many bytes to allocate

data = numpy.array([2., 2., 2., 0.5, 0.4, 0.4, 1.], dtype = numpy.float32).toString()

# Write
buffer.write(0, data, len(data))
# Release buffer
buffer.release()

read方法与write方法类似，它需要buffer.read(offset, data, count)。C++签名被读取（int偏移，空隙*数据，int计数），但是PyQT实现将其抽象为读取（偏移、数据、计数）。p>

它确实假设缓冲区绑定到当前OpenGL上下文