Hi Stuart-
I'm going to move this to idvdevelopers since it's more of an API
question.
Stuart Maclean wrote:
I have a dataset of the form
(lat,lon,TYPE)+
where TYPE can have values say A or B. I would like to plot this
using the idv classes such as LineDrawing, etc, such that the lat,
lon points are joined by lines, but the actual points reflect the A
or B type. Imagine that for types A we want to draw a cross-hair,
whereas for types B we draw a filled circle. I do not want to
distinguish A from B using color, since there are many of these
'vehicle tracks' on one display and the vehicles are already
distinguished using a color per vehicle.
You'd need two Displayables, one for the line and one for plotting
the points. For the lines, you could just create a LineDrawing and
would just need the lat/lon points as a Gridded2DSet with manifold
dimension 1 for the line. You could use MapLines which is a subclass
and probably has some more methods that would be applicable.
For the points, it's a little tricky because the best class for
that is StationModelDisplayable. That would entail you structuring
your data as a FieldImpl of PointObs and creating a StationModel
(layout model) which had a MappedShapeSymbol in it. With the
MappedShapeSymbol, you could have different shapes based on the
value (A or B) (ShapeMap). The StationModel classes are over
in ucar.unidata.ui.symbol.
I must add that the idv api classes are a great step up in
abstraction from the native visad ones. Great job Don and Jeff!
Thanks. The initial work on the DisplayMaster/Displayable classes
was done by Steve Emmerson at Unidata and we've built on that.
Other groups have (e.g., Australian Bureau of Meteorology) have done
other abstractions using "layers" which can encapsulate several
DataReferences.
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
*************************************************************
