[Insight-users] Re: on wrap_itk

Sat Apr 14 16:08:15 EDT 2007

Le 14 avr. 07 à 21:25, anirudh vij a écrit :

> >You didn't sent me the content of your /usr/local/InsightToolkit/
> >WrapITK/Python/itkExtras/ directory. Can you do that ?
>
> this directory contains an __init__.py file.the reference to show3d 
> () in this is:
>
> def show(input, **kargs) :
>   """display an image
>   """
>   import itk
>   img = image(input)
>   if img.GetImageDimension() == 3 and "show3D" in dir(itk):
>           return itk.show3D(input, **kargs)
>   else :
>           # print "2D not supported yet, use the 3D viewer."
>           return show2D(input, **kargs)
>

ItkVtkGlue is not properly installed on your system :-)
Please go back to Wrapping/WrapITK/ExternalProjects/ItkVtkGlue/ and  
reinstall it. Be sure to set the same prefix than the one you used  
for ITK itself.

> >Are you using PyBuffer to pass data from ITK to scipy ?
>
> afraid had'nt heard of it till now.was using the most primitive  
> way,temp files :)
> but will try now for sure.
>
> there's something new thats come up in the meanwhile.the best way  
> to explain would be to paste ipython's output:
>
>
> In [159]: bbox=lab_stat.GetBoundingBox(1)
>
> In [160]: bbox
> Out[160]:  
> '_e0b9c208_p_std__vectorTlong_int_std__allocatorTlong_int_t_t'
>
> In [161]: type(bbox)
> Out[161]: <type 'str'>
>
> here lab_stat is a itk.LabelStatisticsImageFilter instance.why is  
> bounding box being returned as string instead of std::vector?
> is'nt std::vector wrapped by default?the wrap_itk paper mentions  
> that vector_double wrapping is off by default.is that so for  
> vector_long and vector_unsignedshort also?

std::vector and itk::Vector are not the same classes.
Some instantiations of itk::Vector are available, but sadely, there  
is no instantiation of std::vector available with wrapitk.
If you are interested by getting some infos on labelized objects, you  
can install one of my contributions to ITK:

   http://voxel.jouy.inra.fr/darcsweb/darcsweb.cgi? 
r=binaryAttributeMorphology;a=summary

you can get it with darcs (www.darcs.net):

   darcs get http://voxel.jouy.inra.fr/darcs/contrib-itk/ 
binaryAttributeMorphology/

an example which use that contrib can be found at:

   http://voxel.jouy.inra.fr/darcs/agrobi/agrobi.py

The bounding box is provided by the method GetRegion() of the  
LabelObject.
There is not much doc for now I'm afraid.

Gaëtan

>
> >Can you send a short script which reproduce
> >the problem ?
> this problem happens for all filters.i will send the script file in  
> a next post.
>
> On 4/14/07, Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr> wrote:
> Hi,
>
> Please keep the discussions about wrapitk on the user list: it can
> help others to fix there problems in the future :-)
>
> Le 14 avr. 07 à 11:09, anirudh vij a écrit :
>
> > hi gaetan,
> > kudos once again for creating wrap itk.It has speeded up my work a
> > lot.Prototyping is truly horrible in cpp.
> >
> > in course of my use i faced some problems.hope you'll be able to
> > guide towards the solution.
> >
> > 1)still could'nt get show3d() working.my versions of qt and pyqt
> > are both 4.2.3. maybe thats the problem.
>
> it shouldn't prevent you to see the show3D class in the itk module.
>
>
> > anyway,i installed mayavi which has excellent python
> > compatibility.i would like to write a wrapper for show3d() which
> > would make a call to mayavi whenever show3d() is called,maybe with
> > some extra options.where should i place the .py file give that no
> > show3d.py currently exists in /usr/local/lib/InsightToolkit.maybe
> > you could consider moving the itk.show framework entirely to mayavi
> > in the future .it would mean fewer dependencies on one hand,and a
> > unified show2d() and show3d() interface on the other(sice mayavi
> > can show 2d images with extra options of visualization)
>
> I'm still looking for a good interface for python and ITK. I'm not
> sure that mayavi is what I exactly what I want though, but have an
> interface for it would be a great feature :-)
> The best way to go, IMHO, is to create a mayavi function or class to
> display the image, and to put that function in a module in the
> itkExtras dir, so its content will automatically appear in the itk
> module - show3D is made that way.
>
> >
> > 2)whenever a filter is incorrectly initialized and an update is
> > triggered,wrap_itk stops working for that session.subsequent
> > imports produce the error:module swig not defined.
> >
> > 3)after succesfully triggering an update on a filter,if some
> > parameters are changed subsequently,and an update is again
> > triggered,there is no change.for example.
> > dil=itl.BinaryDilateImageFilter.IFUS3US3.New()
> > dil.SetKernel(se)  #se is some strel
> > dil.Update()   #works fine
> > dil.SetKernel(se2)
> > dil.Update() #no change :(
> >
> > is it a problem with my installation or is this the way it is
> > supposed to behave.
>
> it looks like a problem. Can you send a short script which reproduce
> the problem ? It would help a lot to reproduce it and to debug.
>
> >
> > currently i am trying to extend wrap_itk using scipy (i.e. make
> > calls cleaner,add routines that directly do common tasks etc).
>
> Are you using PyBuffer to pass data from ITK to scipy ?
>
> > i will report any major successes.
> >
>
> Please, do :-)
>
> Gaëtan
>
>
> --
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr
>
>
>
>
>

--
Gaëtan Lehmann
Biologie du Développement et de la Reproduction
INRA de Jouy-en-Josas (France)
tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
http://voxel.jouy.inra.fr

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 186 bytes
Desc: =?ISO-8859-1?Q?Ceci_est_une_signature_=E9lectronique_PGP?=
Url : http://public.kitware.com/pipermail/insight-users/attachments/20070414/d8777ad2/PGP.pgp