I understand the role of the ShapeSet better, now. So what I was trying
to do doesn't make sense. In my mind, the ShapeSet values were the
indices of the array of GeometryArrays. Now I see that the ShapeSet
defines a set of range values that are mapped to array indices by
(linear?) interpolation. Is that right?
I was tring to put complex data into a single shape. I guess I'll use a
new shape for each component.
Thanks,
Doug
Don Murray wrote:
>
> Hi Doug-
>
> I can't answer all your questions, but:
>
> Doug Lindholm wrote:
>
> > Is there a case where it makes sense to use anything other than an
> > Integer1DSet for the ShapeSet? It's an extra level of complexity in an
> > already complex API for making shapes.
>
> Here's something Bill sent to me a while back when I was trying
> to figure out how to do station model plots:
>
> >Shouldn't need a custom DataRenderer. If your data have MathType:
> >
> > (index -> (lat, lon, t, td, wx, ...))
> >
> >you'd map:
> >
> > lat -> YAxis
> > lon -> XAxis
> > t -> Shape
> > td -> Shape
> > wx -> Shape
> > . . .
> >
> >In the ShapeControl for t and td, you'd set the quantization of
> >t and td values in the setShapeSet(set) where set might be
> >Linear1DSet(t, -100.0, 100.0, 2001) for 0.1 degree resolution,
> >and you'd use PlotText.render_label() to create the VisADLineArrays
> >for each 0.1 degree from -100.0 to 100.0, all passed to setShapes()
> >(see Test47.java for an example). You might add slight offsets to
> >the 'double[] start' locations for t and td, to get them offset
> >from the center of the station plot. The station plot center
> >position will be determined by the lat and lon mappings.
> >
> >For wx you'd have some numerical codes for each weather symbol,
> >and put the symbols in VisADLineArrays passed to wx's setShape().
> >I think Tom may have started this a long time ago.
> (He hasn't, but he's thinking about it)
>
> I'm not sure what you are trying to do with the Field, but
> I think the idea is that each parameter has it's own ShapeControl
> and the VisADGeometryArrays are defined for each parameter.
> For numbers, it's easy, for symbols there should be a static
> factory for them (cloud cover, weather, barometric tendency, etc).
>
> I agree that the Shape model is complex, but it is also
> very powerful (kinda like the rest of VisAD ;-)). It just
> takes a little while to "get your mind right".
>
> Hope this helps.
>
> Don
> *************************************************************
> Don Murray UCAR Unidata Program
> dmurray@xxxxxxxxxxxxxxxx P.O. Box 3000
> (303) 497-8628 Boulder, CO 80307
> http://www.unidata.ucar.edu/staff/donm
> *************************************************************
--
*----------------------------------------------------------------------*
| Doug Lindholm, Software Engineer | E-mail: lind@xxxxxxxx |
| Research Applications Program | Phone: 303-497-8374 |
| National Center for Atmospheric Research | |
| P.O. Box 3000 | There's no place |
| Boulder, Colorado 80307-3000 | like $HOME |
*----------------------------------------------------------------------*
