GitLab

Situation : when two 1D entities of the microstructure boundary have the same endpoints but only one is actually on the border of the unit cell. The second 1D entity can be an circular arc.
E.g.  `pantograph_RVE(1, 1, 1/3, junction_r=r, offset=(0.05, 0.0))`

Remove all the physical groups from the model except those representing the part.
Reason ? Delete the physical groups that may be used when defining the RVE for splitting/cutting the RVE into several parts.

+ update .gitignore

Handling of simulations without imposed force.

New organisation of functions designed for generating periodic meshes.

New organisation of functions designed for generating periodic meshes.

Possibility to use vertex values in DG elements for kind of interpolation

Use of the function designed to import a mesh with xdmf format.

Gmsh2DPartFromRVE : This function can now be used even when the RVE contains several domains.
pantograph_E11only_RVE : Modified to avoid creating a geometry  with junctions two times thinner on the horizontal borders.

Removing of the corresponding staticmethod from the Gmsh2DRVE class

Handling cases of microstructures that do not intersect the horizontal or vertical edges of the cell.

New cells : only pantographic effect, no shear floppy mode

Links to the new Navier website

+ Split generate_mesh_2D
+ Offset functions (WIP)

Add toolbox_gmsh to __all__
+ Black

Periodic BCs do not seem relevant for DG function spaces. In addition they may not be supported for DG function spaces for now : https://fenicsproject.discourse.group/t/dg-for-advection-equation-with-periodic-boundary-condition/272
Test for demo_homog2D.py : values of Chom and D# have not changed.
+ Black

Xdmf management

See merge request !3

Including subdomains and with plots.
+ minor corrections linked to the msh_conversion function.

with andling of subdomains and facet regions (can be created with physical groups in gmsh);
+ move msh_conversion function to toolbox_gmsh;

Update demo scripts

See merge request !2