VisAD Visualization SpreadSheet
Contents
Introduction
The VisAD Visualization SpreadSheet is a program that utilizes VisAD to
visualize data sets and do simple computations, without forcing the user to do
any Java programming. It is a good example of how to build a user interface
for a VisAD application.
The SpreadSheet is made up of multiple spreadsheet cells, much like
"traditional" spreadsheet programs, except that each cell is capable
of displaying a data object (which can be read from a file or computed from
other cells). To visualize the data object, the SpreadSheet provides an easy
and flexible way to set up ScalarMaps, as well as a formula entry box, for
manipulating data from multiple spreadsheet cells.
Of course, the full flexibility of VisAD is only accessible to programmers
using the class library, but the SpreadSheet is an easy way to visualize
and compute with many data formats, including ASCII, Bio-Rad PIC, BMP, DICOM,
DODS, F2000, FITS, GIF, HDF-5, HDF-EOS, JPEG, LUT, McIDAS, McIDAS ADDE,
netCDF, PGM, PNG, QuickTime, ROI, TIFF, Vis5D, and VisAD binary data files.
Here are some screen-shots of the SpreadSheet in action
(click for a larger picture):
Installation
To run the SpreadSheet, you must have JDK 1.2, available from
Sun's JDK 1.2 web site.
If you want 3-D displays, you'll also need Java3D 1.2.1, which you can obtain
from Sun's Java3D
web site.
You must also have VisAD installed. The
VisAD home page has
detailed information about VisAD, including how to download and install VisAD.
One way to install VisAD is to download
visad.jar,
then include the full path to your visad.jar file in your CLASSPATH environment
variable.
Running the SpreadSheet
Getting started
The SpreadSheet is packaged with VisAD. Once you have set up your visad.jar
file, and modified your CLASSPATH, you can run the SpreadSheet by typing:
java -mx64m visad.ss.SpreadSheet
The '-mx64m' indicates for Java to use up to 64 MB of memory as needed. The
SpreadSheet itself can run without this command flag, but the larger the data
sets that you wish to import, the more memory you'll need.
Importing data
You can import data from a local file, a URL, or an RMI address. To import
data from a file, choose 'Import data' from the 'File' menu, then select the
data file that you want to import using the dialog box.
To import data from a URL, type the URL into the formula bar (the text field
next to the ADD and DEL buttons), then press ENTER.
Warning: VisAD does not yet support loading most file formats through
URLs, making the SpreadSheet unable to import these file types. Only loading
of GIF, JPEG and PNG files through URLs is supported now.
To import data from an RMI address, type the RMI address into the formula bar
with the syntax:
rmi://ip.address/ServName/CellData
where ip.address is the IP address of the RMI server, ServName is the name of
the server, and CellData is the name of the remote data object to import
(e.g., the first data object of spreadsheet cell A1 would be "A1d1"). Local
files can be loaded using the formula bar as well.
In any case, the SpreadSheet will attempt to identify what kind of data it is,
then load it into the current cell. If the data file is loaded successfully,
the SpreadSheet will guess at mappings that might work with the data set. If
the SpreadSheet cannot determine mappings, or you don't like the mappings that
it chooses, you will have to assign them manually.
Setting up mappings from data to display
To manually edit a cell's mappings, choose 'Edit mappings...' from the
'Display' menu.
Here is a screen-shot of the Edit Mappings dialog box, and how to use it:
Using formulas
Sometimes it's nice to be able to compute a data object from other data objects
using a formula. To do this, first select the cell you'd like to assign the
formula to by clicking it, then click within the formula text field, type your
formula in terms of other data objects, and press enter. If the formula
computes successfully, a new data object will appear within the cell.
Otherwise, the cell will contain text listing the errors that occurred during
the evaluation.
For example, say you want to create a data object in cell B2 that is the
average of data objects from cells A1 and B1. You could accomplish this by
highlighting cell B2, then typing:
(A1d1 + B1d1) / 2
in the formula text field, and pressing enter. If the first data object from
cell A1 ("A1d1") and the first data object from cell B1 ("B1d1") have
compatible data structures (MathTypes), then cell B2 will contain a data object
(e.g., "B2d1") that represents their average.
Setting up the SpreadSheet as an RMI server
If you run the SpreadSheet as an RMI server, then SpreadSheets on other
machines will be able to access your SpreadSheet's data objects as RMI
addresses. To run the SpreadSheet as an RMI server, launch the SpreadSheet
with the "-server" flag like:
java -mx64m visad.ss.SpreadSheet -server MyServer
where MyServer is what you wish to name the server. Now other SpreadSheets can
import your SpreadSheet's data objects as explained in the
Importing data section.
Remote collaboration with cloned SpreadSheets
A SpreadSheet can be constructed as a clone of a SpreadSheet RMI server running
on another machine. The cloned sheet will be linked with the server, and
changes to one will be reflected on the other. To create a cloned SpreadSheet,
launch the SpreadSheet with the "-client" flag like:
java -mx64m visad.ss.SpreadSheet -client ip.address/ServName
where ip.address is the IP address of the SpreadSheet RMI server and ServName
is the name of the server.
For more information
See the README.ss file for more detailed information
about the SpreadSheet's features and how to use them.
Loading HDF-EOS and HDF-5 files with the SpreadSheet
Logic to load HDF-EOS and HDF-5 files is currently implemented in VisAD.
However, the loaders must utilize native C code, which is set up to compile
on Sun Solaris or Irix machines only. Native HDF-5 code is available for
several other platforms; see the
JHI5 web site
for more information.
To compile the native code, you will need to take the following steps:
1) Download the
source code
version of VisAD from the VisAD FTP site.
2) Extract the jar file by typing:
jar xvf visad_src-2.0.jar
3) Change your CLASSPATH environment variable to include the parent directory
of the visad directory (instead of the visad.jar file). For example, if you
have installed VisAD to /java/visad, then type:
setenv CLASSPATH /java;.
4) Follow the directions in the visad/data/hdfeos/README.hdfeos and
visad/README files to install HDF-EOS and HDF-5.
5) Set the required HDF-related variables (HDF_INC, HDF_LIB, HDFEOS_INC and
HDFEOS_LIB), to their respective include and library directories of your
HDF installations, if they are not set already.
6) Compile VisAD by typing:
cd visad
make compile
The VisAD source files and all native code will be compiled. Be patient; the
compile can take a few minutes to run depending on your machine. After the
compile finishes successfully, you should be able to load HDF-EOS and HDF-5
files using the VisAD SpreadSheet.
If you are using Windows, you can enable HDF-5 support by installing the
HDF-5 library and
Java HDF5 Interface
according to NCSA's
installation instructions.
Sample data to try with the SpreadSheet
Here is some interesting sample data to try out with the SpreadSheet.
All of these files are in the public domain.
b2rlc.nc (130 KB)
-- An interesting netCDF file with five time steps. Try mapping wnum1
to X Axis, atmosphericRadiance to Y Axis, and
time to Z Axis, Animation, or Select Value.
small.v5d (518 KB)
-- This small Vis5D file has three variables. Try mapping time to
Animation, row to X Axis, col to
Y Axis, lev to Z Axis, and U, V,
and W to Iso-contour. Then, use the contour controls to
change the values of U, V, and W.
rad.nc (133 KB)
-- This netCDF file's default mappings are interesting, but also be sure to
try mapping mdr_x to X Axis, mdr_y to
Y Axis, longitude to Red, latitude to
Green, and mdr to Blue.
rosat.fits (523 KB)
-- Using the default mappings, change the hi value of the range of
value mapped to RGB to around 120, and the image will be much
clearer.
sphere.jpg (5 KB)
-- The default mappings will display this JPEG file fine, but you can spice up
the visualization by mapping ImageLine and ImageElement to
Latitude and Longitude instead of X Axis and
Y Axis, and perhaps even mapping Red, Green, or
Blue to Radius.
The MiniSheet example
The MiniSheet program, found in the visad/examples directory of your VisAD
source file installation, provides a simple example of how to use the
visad.ss classes in an application. Of course, visad.ss.SpreadSheet itself
provides a detailed and rich example, but MiniSheet is much more approachable
because of its simplicity (it is less than 300 lines of code, much of which
sets up the GUI).
Here is a screenshot of the MiniSheet program (click for a larger picture):
Getting help
If you need help using the SpreadSheet or have a question about it, first check
the README.ss file. If you can't find an answer there,
feel free to send e-mail to the
VisAD mailing list.
To subscribe to the VisAD mailing list, send e-mail to
majordomo@ssec.wisc.edu with no
subject and
subscribe visad-list
as the first line in the message body.
To report a bug in the SpreadSheet, please send e-mail to the
VisAD mailing list with a
description of the problem, and if possible, a list of steps to reproduce it.
Keep in mind that the SpreadSheet runs using VisAD functionality, which in
turn takes advantage of JDK 1.2, which runs on top of your operating system,
so any bug you find could be anywhere within this hierarchy. Note especially
that if the SpreadSheet fails to load a data file, chances are that the
SpreadSheet is not at fault, but rather the corresponding VisAD data form is
not yet advanced enough to handle that specific data file.
This page was last updated Wednesday, 27 March, 2002.
