You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment the function write_tikz outputs nodes of the mesh and connects them with straight line segments (codim=1 entities). I propose to add another option to WriteTikz named TikzOutputControl::SecondOrder which will connect the nodes using parabolas.
This functionality is mainly interesting for Geometry Elements that use second order (or higher) geometry approximation in which case the trias/quads can be represented exactly. But Implementing this functionality is not trivial because the Geometry interface doesn't provide us with the parametrization of the parabola directly. I propose to solve this problem as follows:
For every edge connecting two nodes we evaluate the edge geometry element at three points: at the two endpoints and at the point in the middle (average of the endpoints). We then construct a parabola that goes through these three points and give this parametrization to tikz.
The text was updated successfully, but these errors were encountered:
I will implement the same functionality for lf::io::writeMatplotlib. As I am not very well versed in MATLAB, I think it would be easier if you extend lf::io::writeMatlab@hiptmair. What do you think?
anianruoss
changed the title
Adapt write_tikz to support higher order polynomials
Adapt lf::io to support higher order polynomials
Mar 3, 2019
I will take care of the MATLAB visualization tool, but only in the near future, because it will be an afternoon's work to implement and test everything.
At the moment the function
write_tikz
outputs nodes of the mesh and connects them with straight line segments (codim=1 entities). I propose to add another option to WriteTikz namedTikzOutputControl::SecondOrder
which will connect the nodes using parabolas.This functionality is mainly interesting for Geometry Elements that use second order (or higher) geometry approximation in which case the trias/quads can be represented exactly. But Implementing this functionality is not trivial because the
Geometry
interface doesn't provide us with the parametrization of the parabola directly. I propose to solve this problem as follows:For every edge connecting two nodes we evaluate the edge geometry element at three points: at the two endpoints and at the point in the middle (average of the endpoints). We then construct a parabola that goes through these three points and give this parametrization to tikz.
The text was updated successfully, but these errors were encountered: