Re: Scaling of an external TransformGroup

On Sat, 16 Nov 2002, Olav Rybatzki wrote:
> first of all, thank you for the time you spend on my problem. Yes i used
> a Gridded3DSet with manifold dimension = 1, but what i get is not what i
> need. To explain you what the problem is, i have attached a very simple
> 2D example (in 3D it's allways the same). When you look on the picture
> you can see the grid i have to visualize. These are results from CFD
> codes (Computational Fluid Dynamic). The circle is (in this example) the
> geometry that was analyzed in a specific fluid flow )This is the reason
> why no points are defined within the circle, the flow goes arround the
> geometry). In this case each quadrilatral is a computing cell, a so
> called "volume". The connections between the points are given by a
> connecting list where each quadrilateral is defined by 4 indizes. From
> this connecting list it is common to create an IndexedArray. So i think
> this is the easiest way. If there is no easy way to get the scaling from
> VisAD, could you explain me where VisAD set the scaling? Maybe i can
> extract the transform with this information.
> To pass my data into VisAD i rebuild my connecting list, from
> quadrilateral or hexahedron to triangle or tetrahedron, and create a
> DelaunayCustom with it.

You can apply the scaling for ScalarMaps to XAxis, YAxis and
ZAxis by calling the scaleValues() methods of the ScalarMaps.

But if I were you, I'd simply "draw" my lines as a bunch
of GriddedSets with manifold dimension = 1.

> Another question in this context is, are there any limitations in
> displaying streamlines? As i explained, i am using an Irregular3DSet for
>   my data. A few days ago i played arround with Display.FlowXX and
> FlowXControl. The vector plot works, but when i switch to streamline
> nothing happened.

Streamlines only work for GriddedSet domains. If you need to
see Streamlines for a flow FlatField over an IrregularSet
domain, you need to resample() to a GriddedSet (if its high
resolution, the sampling error will be minimized - there is
already implicit interpolation in the stream line calculation.

Good luck,