FerriteViz.jl

FerriteViz.jl is a small package to visualize your Ferrite.jl results. Currently all Makie backends are supported and thus, you can visualize your results in a GLMakie window, inside Pluto/Jupyter notebooks via WGLMakie and produce nice vector graphics with CairoMakie.

In the future this package tries to adapt also other plotting packages, such as Plots.jl and PGFPlotsX.jl. Contributions are highly welcome.

Getting Started

Install FerriteViz.jl with the in-built package manager of Julia

pkg> add FerriteViz

Do your computation with Ferrite.jl and save the used DofHandler and solution vector into a variable. Pass those two variables into the FEData constructor

ds = FEData(dh,u)

Besides the dof fields, an FEData carries named point- and cell-data arrays (FerriteViz.set_point_data!, FerriteViz.set_cell_data!); filters read and write them by name, and every plot can be colored by any of them.

Now, you can use solutionplot, meshplot, arrowplot, surfaceplot or the viewer via ferriteviewer — and compose transformations ParaView-style by piping the data through filters:

solutionplot(ds |> WarpByVector(:u, 2.0) |> Gradient(:u) |> VonMises(); color=:vonMises)

Note that the mutating solutionplot!, meshplot!, arrowplot! and surfaceplot! are available as well.

Unique features

This package offers a set of unique features that are not easily reproducible with other export options of Ferrite.jl:

• a composable, reactive filter pipeline (ParaView's Source → Filter → Representation model): FerriteViz.WarpByVector, FerriteViz.Gradient, FerriteViz.CrinkleClip, FerriteViz.Refine, FerriteViz.FirstOrderRefinement, FerriteViz.ExtractComponent, FerriteViz.Magnitude, FerriteViz.Norm1, FerriteViz.VonMises, FerriteViz.Deviator, FerriteViz.Threshold, FerriteViz.Derive, ...
• FerriteViz.AddQuadraturePointData renders internal variables — data known only at the quadrature points, with no interpolation defining it elsewhere — by partitioning every cell into the exact Voronoi regions of its quadrature points, so the values are neither averaged over the cell nor smoothed onto a nodal field
• FerriteViz.solutionplot FE solution contour plot on arbitrary finite element mesh (in Makie called mesh plots)
• FerriteViz.ferriteviewer viewer with toggles and menus that update the plot, composable through a layout(ds, state) hook and pluggable FerriteViz.Controls
• FerriteViz.meshplot plots the finite element mesh and optionally labels nodes and cells
• FerriteViz.arrowplot - also called quiver plots, in paraview glyph filter
• FerriteViz.surfaceplot 2D solutions in 3D space as surface, in paraview warp by scalar filter
• synchronous plotting while your simulation runs with any of the above listed options
• mutating versions of the above listed functions (except for the viewer)
• deformed plots for any representation via the FerriteViz.WarpByVector filter
• support for custom cell types by implementing a single FerriteViz.reference_tessellation method (walked through on the custom cells page)
• Wedge and Pyramid cells out of the box, and curved (higher-order geometry) cells that tessellate through their geometric interpolation
• full integration into the Makie ecosystem, e.g. themes, layouts etc.
• GPU powered plotting with GLMakie.jl, jupyter/pluto notebook plotting with WGLMakie.jl and vector graphics with CairoMakie.jl
• visualization of high order solutions via first order refinement
• visualization of non-conforming solutions, e.g. for Crouzeix-Raviart ansatz

Viewing the docs locally

To view the docs locally use the provided live server:

include("docs/liveserver.jl")

Opening the html files in the browser directly might fail with a CORS error, manifesting itself figures which don't render correctly.