Update waves

Created by: acrovato

Context

This pull request brings the last developments from the flow module as well as some upgrade to waves.

Major changes in flow

• flow is now able to compute 3D flows
• The Jacobian (tangent) matrix for the Netwton method is now fully analytical
• The adjacent elements identification is now based on ```std::unordered_set````

Major changes in waves

• tbox and fwk are now included in a namespace
• waves is now fully included in a namespace
• All elements now have a private memory (Mem) which can be publicly accessed
• All elements now have a cache including Gauss points (Gauss) which can be publicly accessed
• MshDeformation now uses a linear elasticity solver
• MshCrack class implemented (creates duplicate internal boundaries in meshes)

Test

The ctest battery has been run on gaston, 100% tests passed under 110s.

Future work

• Use automated indentation rules for the whole code
• Move methods using VTK outside of MshData
• Clean tbox (mainly MshData, Tri3, Tera4 and Hex8)

flow/case/polar.py

@@ -42,12 +42,12 @@ def main():
 
         # add a medium "air"
         if p['M_inf'] == 0:
-            pbl.set(f.Medium(msh, p['Fluid'], f.Fun0EleRhoL(), f.Fun0EleMachL(), f.Fun0EleCpL(), f.Fun0PosPhiInf(alpha)))
+            pbl.set(f.Medium(msh, p['Fluid'], f.Fun0EleRhoL(), f.Fun0EleMachL(), f.Fun0EleCpL(), f.Fun0PosPhiInf(p['Dim'], alpha)))
         else:
-            pbl.set(f.Medium(msh, p['Fluid'], f.Fun0EleRho(p['gamma'], p['M_inf'], p['M_crit']), f.Fun0EleMach(p['gamma'], p['M_inf']), f.Fun0EleCp(p['gamma'], p['M_inf']), f.Fun0PosPhiInf(alpha)))
+            pbl.set(f.Medium(msh, p['Fluid'], f.Fun0EleRho(p['gamma'], p['M_inf'], p['M_crit']), f.Fun0EleMach(p['gamma'], p['M_inf']), f.Fun0EleCp(p['gamma'], p['M_inf']), f.Fun0PosPhiInf(p['Dim'], alpha)))
 
         # add initial condition
-        pbl.add(f.Assign(msh,p['Fluid'], f.Fun0PosPhiInf(alpha)),"IC")
+        pbl.add(f.Assign(msh,p['Fluid'], f.Fun0PosPhiInf(p['Dim'], alpha)),"IC")
         # add farfield boundary conditions
         for bnd in p['Farfield']:
             pbl.add(f.Neumann(msh, bnd, tbox.Fct1C(-U_inf[0], -U_inf[1], 0.)))