[vtk-developers] numpy_to_vtk - contiguous array error

[Eric E. Monson]

Hey Randy,

I'm not an expert on this, but the issue is with whether the numpy array is in "C" or "Fortran" order in memory -- numpy_to_vtk() expects "C" ordering. You can see whether the array is "C_CONTIGUOUS" or "F_CONTIGUOUS" by typing array_name.flags

There might be a better way, but you can get around this by making a copy that has the right order, either by invoking .copy() which uses the "C" default, or by explicitly copying with the correct order flag.

-Eric

# =================================
In [1]: import numpy as N

In [2]: from vtk.util import numpy_support as VN

In [3]: xx = N.array([[1,2],[3,4]])

In [4]: xx.flags
Out[4]: 
  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  UPDATEIFCOPY : False

In [5]: VN.numpy_to_vtk(xx)
Out[5]: (vtkLongArray)0x10bea32a0

In [6]: xx = N.array([[1,2],[3,4]], order='F')

In [7]: xx
Out[7]: 
array([[1, 2],
       [3, 4]])

In [8]: VN.numpy_to_vtk(xx)
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)

/Users/emonson/Programming/Python/VTK/Charts/<ipython console> in <module>()

/Users/emonson/Programming/VTK_git/VTK/build/Wrapping/Python/vtk/util/numpy_support.pyc in numpy_to_vtk(num_array, deep, array_type)
    130 
    131     shape = z.shape
--> 132     assert z.flags.contiguous, 'Only contiguous arrays are supported.'
    133     assert len(shape) < 3, \
    134            "Only arrays of dimensionality 2 or lower are allowed!"

AssertionError: Only contiguous arrays are supported.

In [9]: xx.flags
Out[9]: 
  C_CONTIGUOUS : False
  F_CONTIGUOUS : True
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  UPDATEIFCOPY : False

In [10]: yy = N.array(xx, order='C')

In [11]: VN.numpy_to_vtk(yy)
Out[11]: (vtkLongArray)0x10c017270

In [12]: zz = xx.copy()

In [13]: VN.numpy_to_vtk(yy)
Out[13]: (vtkLongArray)0x10c0178d0



On Nov 8, 2010, at 2:27 PM, Randy Heiland wrote:

> Can someone shed light on the following error?
> 
> >>> teMtx
> array([[  0.00000000e+00,   2.21433476e-07,   1.29724158e-06, ...,
>           1.13706443e-06,   1.22732023e-06,   1.80498142e-06],
>        [  3.49344691e-07,  -1.63000845e-07,   4.27805878e-06, ...,
>           5.69157035e-07,   1.44623813e-07,   1.07077057e-07],
>        [  1.12878051e-08,   7.65250526e-06,  -3.56727587e-09, ...,
>           4.44896483e-08,   5.03697289e-07,   2.19654679e-07],
>        ..., 
>        [  9.13691496e-08,   8.34874369e-08,   6.68844738e-07, ...,
>           0.00000000e+00,   4.14336786e-06,   1.54018126e-07],
>        [  1.56768727e-06,   1.26927141e-07,   1.86173239e-06, ...,
>           1.76178961e-06,  -2.94978281e-06,   2.25371326e-06],
>        [  2.13822621e-06,   2.01947558e-08,   6.85551555e-07, ...,
>           3.93482720e-07,   8.21207577e-07,   0.00000000e+00]])
> >>> type(teMtx)
> <type 'numpy.ndarray'>
> >>> teMtx.size
> 106929
> >>> teMtx.shape
> (327, 327)
> >>> vtk_arr2 = numpy_to_vtk(teMtx)
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "<my-path>/Wrapping/Python/vtk/util/numpy_support.py", line 132, in numpy_to_vtk
>     assert z.flags.contiguous, 'Only contiguous arrays are supported.'
> AssertionError: Only contiguous arrays are supported.
> 
> However, the following works fine:
> >>> m2 = numpy.array([[1.0,2.0],[3.0,4.0]])
> >>> type(m2)
> <type 'numpy.ndarray'>
> >>> m2.size
> 4
> >>> m2.shape
> (2, 2)
> >>> vtk_arr_m2 = numpy_to_vtk(m2)
> >>>
> 
> >>> vtk.vtkVersion.GetVTKVersion()
> '5.7.0'
> 
> 
> teMtx happens to come from a 'scipy.io.loadmat', however, via inspection, it appears to be the same as the simpler m2 array.
> 
> thanks, Randy
> 
> _______________________________________________
> Powered by www.kitware.com
> 
> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
> 
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtk-developers
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20101108/b0e33a12/attachment.html>