[Dune] Extract neighbouring element opposite to a certain vertex

[Agnese, Marco]

Hi Carsten, hi Andreas,
yes, you are right. Looking at the vertices coordinates the faces are perfectly defined and everything is perfectly fine (I am looking at both the figures in section 3.). 

I still think that the covered lines should be dashed to have a coherent plot but maybe that's is a matter of personal taste.

Thank you very much for all the helps,
Best,
Marco

________________________________________
From: Carsten Gräser <graeser at mi.fu-berlin.de>
Sent: 25 February 2016 16:44
To: Agnese, Marco; dune at dune-project.org
Subject: Re: [Dune] Extract neighbouring element opposite to a certain vertex

I'm sorry, but you must be looking
at a different picture. The triangle
denoted '0 bottom' is clearly
spanned by the vertices called

0 (0,0,0)
1 (1,0,0)
2 (0,1,0)

And vertex 3 (0,0,1) is clearly on
top, opposite to this face.

Despite the fact that it's grayed out
it may be slightly unclear if the vertex 2
is behind or in front of the screen plain,
but this does not change the vertex
face relation in any way.
Best,
Carsten

Am 25. Februar 2016 17:25:49 MEZ, schrieb "Agnese, Marco" <m.agnese13 at imperial.ac.uk>:
>From the picture it seems that the vertex 3 is opposite to the so
>called front face (which has index 1) and that vertex 2 is opposite to
>the bottom one (index 0). It looks that the bottom face is defined by
>the vertices 0, 1 and 3 and that the vertex 2 is outside the plane of
>the screen. Obviously if the bottom face is defined by the vertices 0,
>1 and 2 everything works fine. But since there aren't any dashed lines,
>it means that all edges are visible and necessary 2 is the "top''of the
>pyramid pointing outside the screen.
>
>
>Or the fading numbers are the one covered by the faces? In that case
>everything is correct. But this isn't a very standard notation, at
>least for me, so, looking at the picture it comes automatically to me
>the first configuration which is wrong.
>
>I don't know if it will be formatted correctly in the mailing list but
>the reference element should be drawn like this:
>
>
>           3
>         ,/|`\
>       ,/  |  `\
>     ,/    '.   `\
>   ,/       |     `\
> ,/         |       `\
>0-----------'.-----2
> `\.         |      ,/
>    `\.      |    ,/
>       `\.   '. ,/
>          `\. |/
>             `1
>
>
>Hope to have been sufficiently clear :)
>Cheers,
>Maco
>
>________________________________________
>From: Carsten Gräser <graeser at mi.fu-berlin.de>
>Sent: 25 February 2016 16:04
>To: Agnese, Marco; dune at dune-project.org
>Subject: Re: [Dune] Extract neighbouring element opposite to a certain
>vertex
>
>Am 25.02.2016 um 15:59 schrieb Agnese, Marco:
>> Hi Martin,
>>
>>
>> yes I am using conforming meshes.
>>
>>
>> By the way
>>
>https://www.dune-project.org/doc-2.2.1/doxygen/html/group__GeometryGenericReferenceElements.html
>seems
>> wrong for the 3d simplex (section 3.). Shouldn't be 1 bottom and 0
>front
>> in the face numbering picture? Otherwise vertex_idx+oppsite_face_idx
>!=
>> entity_dimension for vertex_idx=3 or 2.
>I can't see any problem here, it's
>
>  3(top vertex) + 0(bottom face) = 3
>
>and
>
>  2(back vertex) + 1 (front face) = 3,
>
>exactly as stated by Martin.
>
>Best,
>Carsten
>
>>
>>
>> Thank you very much,
>>
>> Marco
>>
>>
>>
>>
>------------------------------------------------------------------------
>> *From:* Martin Nolte <nolte at mathematik.uni-freiburg.de>
>> *Sent:* 25 February 2016 14:41
>> *To:* dune at dune-project.org
>> *Subject:* Re: [Dune] Extract neighbouring element opposite to a
>certain
>> vertex
>>
>> Hi Marco,
>>
>> yes, you have to iterate over all intersections to find the
>corresponding
>> neighboring entity (I assume you are using a conforming grid,
>otherwise the
>> neighbor will not be unique).
>>
>> You seek the intersection satisfying
>>
>> indexInIndide() == Entity::dimension - vtxIdx.
>>
>> Note that, in general, the world dimension does not coincide with the
>> dimension (though I guess it does in your case).
>>
>> As to your second question: There is no guarantee on the order the
>faces
>> are
>> traversed in.
>>
>> Best,
>>
>> Martin
>>
>> On 02/25/2016 03:29 PM, Agnese, Marco wrote:
>>> Hi,
>>>
>>> I have a simplex entity and I want to extract the neighbouring
>element
>>> opposite to the vertex with local index vtxIdx.
>>>
>>> Therefore I need to find the intersection opposite to this vertex.
>>>
>>> This intersection has local index faceIdx=worlddim-vtxIdx. Is it
>right?
>>>
>>>
>>> So if I extract entity.template subEntity<1>(faceIdx) I should be
>able to
>>> extract this face. Am I correct? But the face extracted in this way
>is not an
>>> intersection therefore I can't use it to find the neighbouring
>entity. Is
>>> there a solution here?
>>>
>>>
>>> The other solution is to loop over the intersections and extract the
>vtxIdx-th
>>> one. But does it work always? Or which is the ordering of this
>intersections?
>>>
>>> I would expect to have them sorted accordingly the face local index
>(so
>>> faceIdx) but that seems not to be the case. What am I missing?
>>>
>>>
>>> Thank you very much for your help,
>>>
>>> cheers,
>>>
>>> Marco.

--
Prof. Dr. Carsten Gräser
Freie Universität Berlin
Institut für Mathematik
Arnimallee 6
14195 Berlin, Germany
phone: +49 30 838 72637
fax  : +49 30 838 472637
email: graeser at mi.fu-berlin.de
URL  : http://page.mi.fu-berlin.de/graeser