[Dune] UG and DGF with element parameters

[Nicolas Schwenck]

You are absolutely right: At the moment reading parameters also does not 
work with UG and vertices (even though my colleagues swear that they 
have used this functionality some time ago).

Attached you'll find the patch for the insertionIndex methods for 
vertices and elements. Unfortunately I was not able to declare the 
specialized class GridFactory<UGGrid<dimworld> > as a friend so that I 
had to do it for the base class to use the private member target_ in 
ugentity.

Best,
Nicolas

On 08/28/2012 09:27 PM, Dedner, Andreas wrote:
> Well if there is any magic then its not the dgf side thats for sure:
> dgfug.hh:
>     /** \brief return parameter for vertex */
>      std::vector< double > &parameter ( const typename Grid::template Codim< dimension >::Entity &vertex )
>      {
>        if( numParameters< dimension >() <= 0 )
>        {
>          DUNE_THROW( InvalidStateException,
>                      "Calling DGFGridFactory::parameter is only allowed if there are parameters." );
>        }
>        return dgf_.vtxParams[ factory_.insertionIndex( vertex ) ];
>      }
> So I guess that should not work...
> ________________________________________
> From: dune-bounces+a.s.dedner=warwick.ac.uk at dune-project.org [dune-bounces+a.s.dedner=warwick.ac.uk at dune-project.org] on behalf of Oliver Sander [sander at mi.fu-berlin.de]
> Sent: 28 August 2012 19:49
> To: dune at dune-project.org
> Subject: Re: [Dune] UG and DGF with element parameters
>
> Hi Nicolas,
>
>   > Using a dgf file with elementwise parameters and a UG grid throws me
> the aforementioned error so that I assumed that I do not need the
> insertionIndex methods directly but it is needed to handle elementwise
> parameters. I actually don't know why it is not needed for vertex
> parameters.
>
> Maybe that's a bug in the dgf reader?  Or there is some magic default
> implementation
> for the vertices?  I do not know.
>> Anyhow I want to implement the missing methods. I followed the
>> implementation for Alberta and ALU. Where I begin to struggle is: What
>> is the method in UG  (in Alberta it is somehow integrated/hidden in
>> elementInfo() and in ALU it is getIndex() ) which gives me the index
>> for an entity which I can then forward (because you said there will be
>> simply no re-ordering)? Could you please help me on this one?
>>
> Have a look at the implementation of UGGridLevelIndexSet.  There you'll
> find how to extract
> indices from UG.
>
> best,
> Oliver
>
>> Best,
>> Nicolas
>>
>> On 08/21/2012 06:07 PM, Oliver Sander wrote:
>>> Hi Nicolas,
>>> I never implemented that insertion order stuff for UG, because I could
>>> convince UG not ro reorder elements and vertices to begin with, and
>>> because I have so far not met anybody who actually needed those methods.
>>>
>>> If you do need those methods I would ask you to implement them and send
>>> me a patch.  Implementation is trivial: there is simply no reordering.
>>>
>>> best,
>>> Oliver
>>>
>>> Am 16.08.2012 15:03, schrieb Nicolas Schwenck:
>>>> Dear DUNE,
>>>>
>>>> Running the testug in dune-grid/dune/grid/io/file/dgfparser/test with
>>>> the file dune-grid/doc/grids/dgf/examplegrid10a.dgf on the trunk breaks
>>>> for me with
>>>>
>>>> NotImplemented
>>>> [insertionIndex:../../../../../../dune/grid/common/gridfactory.hh:181]:
>>>> insertion indices have not yet been implemented.
>>>>
>>>> I assume that there is a problem with elementwise parameters for UG
>>>> grids. Is this a bug or is it just not yet implemented as the error
>>>> suggests? Does anybody uses ug for dgfs with elementwise parameters and
>>>> can throw me a hint?
>>>>
>>>> Best,
>>>> Nicolas
>>>>
>>>>
>>> _______________________________________________
>>> Dune mailing list
>>> Dune at dune-project.org
>>> http://lists.dune-project.org/mailman/listinfo/dune
>>>
>>
>
> --
> ************************************************************************
> * Oliver Sander                ** email: sander at mi.fu-berlin.de        *
> * Freie Universität Berlin     ** phone: + 49 (30) 838 75348           *
> * Institut für Mathematik      ** URL  : page.mi.fu-berlin.de/~sander  *
> * Arnimallee 6                 ** -------------------------------------*
> * 14195 Berlin, Germany        ** Member of MATHEON (www.matheon.de)   *
> ************************************************************************
>
>
> _______________________________________________
> Dune mailing list
> Dune at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune
>
>
>
> _______________________________________________
> Dune mailing list
> Dune at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune


-- 
_____________________________________________________________________

Dipl.-Ing. Nicolas Schwenck
				
University of Stuttgart
Department of Hydromechanics and Modelling of Hydrosystems

Pfaffenwaldring 61                           phone: +49 711 685 64736
GER-70569 Stuttgart                          fax:   +49 711 685 60430

             email: nicolas.schwenck at iws.uni-stuttgart.de
                 url: www.hydrosys.uni-stuttgart.de
_____________________________________________________________________

-------------- next part --------------
A non-text attachment was scrubbed...
Name: uggrid.patch
Type: text/x-patch
Size: 1752 bytes
Desc: not available
URL: <https://lists.dune-project.org/pipermail/dune/attachments/20120829/cd5526f0/attachment.bin>