Re: [Getfem-users] visitor for points of a mesh region

[Yves Renard]

Dear Tsai,

Your code I return the dofs which corresponding shape function have a nonzero mean on the region (but be carefull, some shape functions of the P2 element in 2D have a zero mean). If you need the dof on a region, there exist a method of the mesh_fem object with return them. If you whant the points indices (with the numeration correpsonding to the mesh, not to a mesh_fem) the only mean, I think, is to iterate on the elements and to use the methods  ind_point_of_convex(..) and  ind_points_of_face_of_convex of the mesh object.


Your code III seems to me correct. What kind of problem do you have ?

Yves.


Le 10/07/2012 03:19, CC Tsai a écrit :

 

Dear Yves,

 

Sorry for my disturb. Right now, I can accomplish what I want by the code I in the last of the email.

 

Second, you have memtioned that do not use the copy constructor for getfem::mesh. Then, what will you suggest for the situation in my code II? I tried the code III which however has some problems.

 

sincerely

Tsai 

 

code I:

 

dal::bit_vector MeshRegion_points_index(const getfem::mesh_fem& meshFem,const getfem::mesh_im& meshIm,unsigned int rg_id)
 {
  unsigned int nb_dof=meshFem.nb_dof();
  dal::bit_vector temp;
  vector<REAL> Btemp(nb_dof);
  gmm::resize(Btemp,nb_dof);  gmm::clear(Btemp);

  getfem::generic_assembly assem;
  assem.push_mi(meshIm);
  assem.push_mf(meshFem);
  assem.push_vec(Btemp);
  assem.set("V(#1)+=comp(Base(#1))(:)");
  assem.assembly(rg_id);

  for(unsigned int i=0;i<nb_dof;i++)
   if(Btemp[i])temp[i]=true;

  return temp;
 }

code II:

 

class FemModel

{

 mesh msh; 

 mesh_fem mf;

 mesh_im mim;

 FemModel(mesh MSH,mesh_fem MF, mesh_im MIM):msh(MSH),mf(MF),min(MIM){}

}

 

code III:

 

class FemModel

{

 mesh msh; 

 mesh_fem mf;

 mesh_im mim;

 FemModel(mesh MSH,mesh_fem MF, mesh_im MIM) {

  msh.copy_from(MSH);
  mim.init_with_mesh(msh);
  mf_u.init_with_mesh(msh);

 }

}

 

_______________________________________________
Getfem-users mailing list
address@hidden
https://mail.gna.org/listinfo/getfem-users

-- 

  Yves Renard (address@hidden)       tel : (33) 04.72.43.87.08
  Pole de Mathematiques, INSA-Lyon             fax : (33) 04.72.43.85.29
  20, rue Albert Einstein
  69621 Villeurbanne Cedex, FRANCE
  http://math.univ-lyon1.fr/~renard

---------