[vtk-developers] Python wrapper change summary

Fri Aug 6 15:52:04 EDT 2010

Hi Folks,

I have few remarks regarding the PyQt and SIP that seems to be used to wrap
QUISupport/Qt. See
http://vtk.org/gitweb?p=VTK.git;a=history;f=GUISupport/Qt;h=d9ee2891fdb603124f3d8928f972a5252353e423;hb=HEAD

SIP :

"SIP is licensed under similar terms as Python itself. SIP is also licensed
under the GPL (both v2 and v3). It is your choice as to which license you
use. If you choose the GPL then any bindings you create must be distributed
under the terms of the GPL."
Source: See
http://www.riverbankcomputing.co.uk/static/Docs/sip4/introduction.html#license

PyQt:

"PyQt, unlike Qt, is not available under the LGPL. If you use the GPL
version of PyQt then any code you develop with it must be distributed under
a compatable license."
Source: http://www.riverbankcomputing.co.uk/software/pyqt/license

Both SIP and PyQT are GPL. If I understand properly, we can't use them in
VTK. Am I misunderstanding something ?

That's why PySide project was created:

"The PySide project provides LGPL-licensed Python bindings for the Qt
cross-platform application and UI framework. PySide Qt bindings allow both
free open source and proprietary software development and ultimately aim to
support all of the platforms as Qt itself."

In the same time, in the context of CTK, I am working with PythonQt and
PythonQtGenerator. Both are now cmake-ified. I am also working on a set of
macro similar to vtkWrapPython.

The macro ctkMacroWrapPythonQt will allow to either do a *full* or a
*light*wrapping.

The light wrapping will provide the ability to create and delete object
(derving from QObject) in python and also call slot and methods added using
Q_PROPERTY. To achieve such things, a small wrapper allowing to decorates
the constructor and destructor is created.

The full wrapping will rely on PythonQtGenerator (a modified version of the
generator used by QtScript) and will allow to have a full wrapping of enums,
virtual method, non-QObject, ...

Will keep you posted.

Thks
Jc

On Fri, Aug 6, 2010 at 3:08 PM, Marcus D. Hanwell <
marcus.hanwell at kitware.com> wrote:

> On Fri, Aug 6, 2010 at 3:01 PM, David Gobbi <david.gobbi at gmail.com> wrote:
>
>> Hi All,
>>
>> This is a brief summary of new vtk/python features for people to try:
>>
>> 1) The wrapping of vtkVariant, vtkArrayRange, vtkArrayCoordinates etc.
>> has changed.  Each of these types now has its own "type object" which
>> doubles as the object constructor.  This means that if you do
>> "help(vtkVariant)" in python, all the methods will be listed.  In the
>> future, this also means that I will be able to wrap more operator
>> methods.  For people who want to try out vtkVariant, there is an
>> example here: http://www.vtk.org/Wiki/VTK/Examples/Python/Variant
>>
>> 2) Constants are automatically wrapped, as long as they are defined in
>> one of the following ways in the headers:
>>
>> #define VTK_CELL_SIZE 512
>> const int VTK_CELL_SIZE = 512;
>> enum { VTK_CELL_SIZE = 512, ... };
>>
>> Some important header files aren't wrapped yet, such as vtkCellType.h,
>> but they will be soon.  Enums and constant variables that are class
>> members are also wrapped, so vtkDataObject.POINT_DATA_FIELD can be
>> used.  Note that the vast majority of these enums have BTX/ETX around
>> them and are only wrapped because in the git head I am forcing the
>> wrappers to ignore BTX/ETX in the main "kit" directories.  So the
>> wrapping of these enums will only become permanent when the BTX/ETX
>> markers are removed.
>>
>
> This is great news, and I am very pleased we can start using the enum
> values. Can we start using the enum types in our API too, i.e. go from void
> SetType(int enumValue) to void SetType(enumType enumValue)?
>
>>
>> 3) The vtkDataArray GetTuple() and SetTuple() methods are wrapped.  So
>> python users don't have to use the SetTuple2(), GetTuple3() methods
>> anymore.
>>
>
> Does this mean we could deprecate some of this API, or is it still very
> much required for the Tcl and Java wrappers?
>
>>
>> v = vtkIntArray()
>> v.SetNumberOfComponents(3)
>> v.InsertNextTuple((1, 2, 3))
>> or to insert without intermediate conversion to "double":
>> v.InsertNextTupleValue((1,2,3))
>> (x,y,z) = v.GetTuple(0)
>> or use RHS get method:
>> rval  = [0,0,0]
>> v.GetTupleValue(1, rval)
>>
>> 4) Clinton has added a two-way VTK/Python/SIP bridge and has wrapped
>> the GUISupport/Qt directory.  I've compiled this but haven't tried it
>> yet.  A couple tests/examples would be nice (hint hint).
>>
>> So, that's it for now.  Enjoy.
>>
>
> Sounds great, well worth a few of the length rebuilds required ;-) I will
> be testing some of this out, and would like to start streamlining the API in
> some of the classes I work on.
>
> Thanks!
>
> Marcus
> --
> Marcus D. Hanwell, Ph.D.
> R&D Engineer, Kitware Inc.
> (518) 881-4937
>
> _______________________________________________
> 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
>
>
>

-- 
Phone: 1-518-836-2174
Ext: 304
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20100806/662ed13e/attachment.html>