GitHub

Plotting

Plotting lives in a package extension so the core package never compiles the Makie stack. Load any Makie backend to activate it:

using TwoDG
using CairoMakie   # or GLMakie

mesh = mkmesh_square(17, 17, 3, 0, 1)
sol  = solve(CGProblem(PoissonEquation(), mesh;
                       source = (x, y) -> 2π^2 * sin(π * x) * sin(π * y)))

scaplot(mesh, sol.u[:, 1, :]; show_mesh = true)

Without a backend loaded, the functions error with a load hint.

  • scaplot(mesh, c; limits=nothing, show_mesh=false, figure_size=(800, 800), title="", cmap=:turbo) — filled contour plot of a scalar field c (npl, nt) in DG numbering (curved elements rendered curved). Returns the Makie figure, so it composes with save("plot.png", fig) and animation loops.
  • meshplot(mesh; nodes=false, annotate="") — wireframe of the triangulation; nodes=true marks the DG nodes, and annotate may contain 'p' (number the vertices) and/or 't' (number the elements).
  • meshplot_curved(mesh; nodes=false, annotate="", pplot=0) — wireframe following the curved (isoparametric) element edges, subdividing each element with order pplot for display; useful for checking boundary projection.

All solution objects store their mesh, so the pattern is always scaplot(sol.prob.mesh, sol.u[:, component, :]). For time series, call scaplot inside a loop over saved states (see examples/dg/runeulerchannel_animation.jl for a record-based movie).

The Makie plots are 2D-only. 3D solutions go to ParaView via save_vtk (WriteVTK package extension) — see 3D in TwoDG.