[Insight-developers] Compare text (vtk) files

[Bill Lorensen]

You read the files and compare the # of polys, points, etc. to a
baseline. At least that will tell you if something major changed. You
might even be able to check the topology (face connectivity).

Bill

On Sun, Jul 11, 2010 at 11:05 AM, Dan Mueller <dan.muel at gmail.com> wrote:
> Thanks for your prompt responses! I made all the screenshots for the
> article using paraview :P
>
> I was looking for something a little more automated to validate the
> output. I mean, if I make a change to the code, the only way I can
> verify the change is to open the 20+ vtk output files one by one in
> paraview. That will have to do for now I guess.
>
> Regards, Dan
>
> On 11 July 2010 16:57, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>> You should at least see if it loads into paraview...
>>
>>
>> On Sun, Jul 11, 2010 at 10:53 AM, Dan Mueller <dan.muel at gmail.com> wrote:
>>> Fair enough.
>>>
>>> I see Testing/Review/itkVTKPolyDataWriterTest01 just writes a vtk file
>>> without validating the output either. I guess this is better than
>>> nothing, though the test could be writing complete garbage and still
>>> pass...
>>>
>>> Thanks. Dan.
>>>
>>> On 11 July 2010 16:40, Bill Lorensen <bill.lorensen at gmail.com> wrote:
>>>> Not that I know of. Sounds difficult to me because of file formats
>>>> differences and floating point precision differences on different
>>>> platforms.
>>>>
>>>> Bill
>>>>
>>>> On Sun, Jul 11, 2010 at 10:29 AM, Dan Mueller <dan.muel at gmail.com> wrote:
>>>>> Hi Insight Developers,
>>>>>
>>>>> I am preparing a submission for the Insight Journal. My new filter
>>>>> itk::CuberilleImageToMeshFilter outputs mesh (vtk) files. I want to
>>>>> write some tests which compare the resultant text (vtk) files to
>>>>> baseline files.
>>>>>
>>>>> I know of the ImageCompare tool which can be used to compare output
>>>>> images to a baseline image. Is there a similar tool or CMake command
>>>>> to compare output text (vtk) files to a baseline text file?
>>>>>
>>>>> Thanks in advance.
>>>>>
>>>>> Regards, Dan
>