I need to create variable with type cl.cltypes.uint2 for pyopencl in Python.
Now i have created it this way:
key = np.array([(0x01020304, 0x05060708)], dtype=cl.cltypes.uint2)[0]
Its definitely dirty hack( How to create it with more clean way?
this: key = cl.cltypes.uint2((0x01020304, 0x05060708))
not works because of error: 'numpy.dtype' object is not callable
A quick read of your link suggests that it is making a compound dtype. With out loading and running it, I think your example is something like
In [164]: dt = np.dtype([('x',np.uint16),('y',np.uint16)])
In [165]: np.array([(0x01020304, 0x05060708)], dtype=dt)
Out[165]: array([(772, 1800)], dtype=[('x', '<u2'), ('y', '<u2')])
In [166]: dt((0x01020304, 0x05060708))
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-166-d71cce4777b9> in <module>
----> 1 dt((0x01020304, 0x05060708))
TypeError: 'numpy.dtype' object is not callable
and pulling out one record from the array:
In [167]: np.array([(0x01020304, 0x05060708)], dtype=dt)[0]
Out[167]: (772, 1800)
In [168]: _.dtype
Out[168]: dtype([('x', '<u2'), ('y', '<u2')])
A compound dtype is never callable.
I think a 0d, 'scalar' array is better than an object created with the dtype function (though they have similar methods).
For a compound dtype:
In [228]: v = np.array((0x01020304, 0x05060708), dtype=dt)
In [229]: v
Out[229]: array((772, 1800), dtype=[('x', '<u2'), ('y', '<u2')])
In [230]: type(v)
Out[230]: numpy.ndarray
In [231]: v[()]
Out[231]: (772, 1800)
In [232]: type(_)
Out[232]: numpy.void
In [233]: _231.dtype
Out[233]: dtype([('x', '<u2'), ('y', '<u2')])
You can cast such an array to recarray and get a record object, but I don't think creating these is any easier.
In [234]: v.view(np.recarray)
Out[234]:
rec.array((772, 1800),
dtype=[('x', '<u2'), ('y', '<u2')])
In [235]: _.x
Out[235]: array(772, dtype=uint16)
In [238]: v.view(np.recarray)[()]
Out[238]: (772, 1800)
In [239]: type(_)
Out[239]: numpy.record