Implement FaceQuadratureRule for RefPrism and RefPyramid

[AbdAlazezAhmed]

Closes #772
Changes the transformations used in create_face_quad_rule to use the ones in FerriteViz as they're beneficial for optimizing #743 and IIUC make more sense in general as they respect the normals directions.
Also, the transformation is separated into its own functions such that they can be reused as in #743.
Edit: test coverage should improve with interface test, because it covers face->cell transformations

[codecov-commenter]

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.71% 🎉

Comparison is base (d474caf) 91.88% compared to head (73650a3) 92.59%.
Report is 6 commits behind head on master.

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #779      +/-   ##
==========================================
+ Coverage   91.88%   92.59%   +0.71%     
==========================================
  Files          33       33              
  Lines        4916     4942      +26     
==========================================
+ Hits         4517     4576      +59     
+ Misses        399      366      -33     

Files Changed	Coverage Δ
src/FEValues/face_integrals.jl	100.00% <100.00%> (+5.43%)	⬆️
src/Grid/grid.jl	92.36% <100.00%> (-0.07%)	⬇️
src/Quadrature/quadrature.jl	97.14% <100.00%> (+11.70%)	⬆️

... and 2 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[KnutAM]

Cool!
I just took a quick look through and shared some thoughts (but couldn't look longer as my laptop is running out of battery:)

[fredrikekre]

transfer_point_face_to_cell could perhaps be called reference_face_to_face or something?

[lijas]

Nice work on this PR. Before this PR, the spatial coordinate evaluated on "cell-level" was not always the same as the coordinate evaluated on "face level" In other words, we just took the parametric coordinate on e.g. on a RefLine (say \xi=[0.123]) and transformed in to the Quadrilateral (say to local faceid 4 \xi = [-1, 0.123), however, this would not always correspond to the same spatial coordiante due to the permutation of the face-nodes. This PR seems to fix that.

Is it worth adding test for this (I can do it in that case)?

Somethink like:

using Ferrite

grid = generate_grid(Ferrite.Pyramid, (3,3,3))
dim = Ferrite.getdim(grid)

cellid = 3
cell = grid.cells[cellid]
ccoords = getcoordinates(grid, cellid)
Vec_t = Vec{3,Float64}
Vec_face_t = Vec{2,Float64}

lfaceid = 2
facenodes = Ferrite.faces(cell)[lfaceid]
fcoords = zeros(Vec_t, length(facenodes))
for (i,nodeid) in enumerate(facenodes)
    x = grid.nodes[nodeid].x
    fcoords[i] = x
end

refshape =  Ferrite.getrefshape(cell)
ipcell = Lagrange{refshape,1}()
ipface = Lagrange{RefTriangle,1}()

ξface = rand(Vec_face_t)/4
ξcell = Ferrite.face_to_element_transformation(ξface, refshape, lfaceid)

xface = zero(Vec_t)
for i in 1:getnbasefunctions(ipface)
    xface += Ferrite.shape_value(ipface, ξface, i) * fcoords[i]
end

xcell = zero(Vec_t)
for i in 1:getnbasefunctions(ipcell)
    xcell += shape_value(ipcell, ξcell, i) * ccoords[i]
end

@test xcell == xface

Edit: I noticed that you mentioned this in #743 "syncing/transforming quadrature points"

[AbdAlazezAhmed]

Thanks @lijas ! I'll work on adding the test 😄