ptolemy.plot.compat
Class PxgraphParser

java.lang.Object
  ptolemy.plot.compat.PxgraphParser

public class PxgraphParser
extends java.lang.Object

This class provides backwards compatibility with an older plotting program, pxgraph. It provides two methods, one for parsing command-line arguments, and one for reading binary data from a file. In pxgraph, the binary files have no format information; all format information is provided by command line arguments.

Below we describe the pxgraph arguments. The text is based on the xgraph Unix man page written by David Harrison (University of California). To see the command line options, you can type pxgraph -help.

The pxgraph program draws a graph on a display given data read from either data files or from standard input if one of the arguments is a dash. It can display up to 64 independent data sets using different colors and/or line styles for each set. It annotates the graph with a title, axis labels, grid lines or tick marks, grid labels, and a legend. There are options to control the appearance of most components of the graph.

The input format is similar to graph(1G) but differs slightly. The data consists of a number of data sets. Data sets are separated by a blank line. A new data set is also assumed at the start of each input file. A data set consists of an ordered list of points of the form directive X Y. The directive is either draw or move and can be omitted (Note that with binary data files, you must have a directive, the above statement only applies to ascii format data files). If the directive is draw, a line will be drawn between the previous point and the current point (if a line graph is chosen). Specifying a move directive tells xgraph not to draw a line between the points. If the directive is omitted, draw is assumed for all points in a data set except the first point where move is assumed. The move directive is used most often to allow discontinuous data in a data set. After pxgraph has read the data, it will create a new window to graphically display the data. Once the window has been opened, all of the data sets will be displayed graphically (subject to the options explained below) with a legend in the upper right corner of the screen. To zoom in on a portion of the graph, depress a mouse button in the window and sweep out a region. pxgraph will then the window will be redrawn with just that portion of the graph. pxgraph also presents four control buttons in the lower left corner of each window: Exit, Print, HTML and About.

The Exit button will exit the process. You can also type Control-D, Control-C or q to exit.

The Print button brings up a print dialog window.

The About button brings up a message about pxgraph.

The HTML button prints an HTML file to stdout that can be used to display the file with applet Plot classes (Experimental).

pxgraph accepts a large number of command line options. A list of these options is given below.

=WxH+X+Y

Specifies the initial size and location of the pxgraph window.

-<digit> <name>

These options specify the data set name for the corresponding data set. The digit should be in the range 0 to 63. This name will be used in the legend.

-bar

Specifies that vertical bars should be drawn from the data points to a base point which can be specified with -brb. Usually, the -nl flag is used with this option. The point itself is located at the center of the bar.

-bb

Draw a bounding box around the data region. This is very useful if you prefer to see tick marks rather than grid lines (see -tk). Ignored in the Java version because the plotting area is a different color than the border where the axes are labeled.

-bd <color>

This specifies the border color of the pxgraph window. Unsupported in the Java version.

-bg <color>

Background color of the area where the labels and legend are rendered. In the Java version, this argument takes hexadecimal color values (fffff), not color names. Note that the background of the data plotting region is always white because the dataset colors were designed for a white background.

-bigendian

Data files are in big-endian, or network binary format. See the -binary command line argument documentation below for details about the format. If you are on a little-endian machine, such as a machine with an Intel x86 chip, and you would like to read a binary format file, created on a big-endian machine, such as a Sun SPARC, use the -bigendian flag.

-binary

Data files are in a binary format. The endian-ism of the data depends on which of the two subformats below are chosen. The -binary argument is the primary difference between xgraph and pxgraph. The Ptolemy Project software makes extensive use of -binary.
There are two binary formats, both of which use 4 byte floats.

1. If the first byte of the data file is not a d, then we assume that the file contains 4 byte floats in big-endian ordering with no plot commands.
2. If the first byte of the data file is a d, then we assume that the plot commands are encoded as single characters, and the numeric data is a 4 byte float encoded in the native endian format of the machine that the java interpreter is running on.
   The commands are encoded as follows:

   d <4byte float> <4byte float>

   Draw a X, Y point

   e

   End of dataset

   n <dataset name>\n

   New dataset name, ends in \n

   m <4byte float> <4byte float>

   Move to a X, Y point.

To view a binary plot file under unix, we can use the od command. Note that the first character is a d followed by eight bytes of data consisting of two floats of four bytes.

 cxh@carson 324% od -c data/integrator1.plt
 0000000   d  \0  \0  \0  \0  \0  \0  \0  \0   d   ? 200  \0  \0   ? 200
 0000020  \0  \0   d   @  \0  \0  \0   @   , 314 315   d   @   @  \0  \0
 

For further information about endian-ism, see the -bigendian and -littleendian command line argument documentation.

-brb <base>

This specifies the base for a bar graph. By default, the base is zero. Unsupported in the Java version.

-brw <width>

This specifies the width of bars in a bar graph. The amount is specified in the user units. By default, a bar one pixel wide is drawn.

-bw <size>

Border width (in pixels) of the pxgraph window. Unsupported in the Java version.

-db

Causes xgraph to run in synchronous mode and prints out the values of all known defaults.

-fg <color>

Foreground color. This color is used to draw all text and the normal grid lines in the window. In the Java version, this argument takes hexadecimal color values (fffff), not color names.

-gw

Width, in pixels, of normal grid lines. Unsupported in the Java version.

-gs

Line style pattern of normal grid lines.

-impulses

Draw a line from any plotted point down to the x axis. (This argument is not present in the X11 pxgraph, but it is similar to -nl -bar).

-lf <fontname>

Label font. All axis labels and grid labels are drawn using this font. Note that the Java version does not use X11 style font specification. In the Java version, fonts may be specified as

fontname, where fontname is one of helvetica, TimesRoman, Courier, Dialog, DialogInput, ZapfDingbats.

fontname-style, where style is one of PLAIN, ITALIC, BOLD, i.e. helvetica-ITALIC

fontname-size, or

fontname-style-size, where size is an integer font size in points.

The default is helvetica-PLAIN-12.

-littleendian

Data files are in little-endian, or x86 binary format. See the -binary command line argument documentation above for details about the format. If you are on a big-endian machine, such as a Sun Sparc, and you would like to read a binary format file created on a little-endian machine, such as Intel x86 machine, then use the -littleendian flag.

-lnx

Specifies a logarithmic X axis. Grid labels represent powers of ten. If -lnx is present, then x values must be greater than zero.

-lny

Specifies a logarithmic Y axis. Grid labels represent powers of ten. If -lny is present, then y values must be greater than zero.

-lw width

Specifies the width of the data lines in pixels. The default is zero. Unsupported in the Java version.

-lx <xl,xh>

This option limits the range of the X axis to the specified interval. This (along with -ly) can be used to zoom in on a particularly interesting portion of a larger graph.

-ly <yl,yh>

This option limits the range of the Y axis to the specified interval.

-m

Mark each data point with a distinctive marker. There are eight distinctive markers used by xgraph. These markers are assigned uniquely to each different line style on black and white machines and varies with each color on color machines.

-M

Similar to -m but markers are assigned uniquely to each eight consecutive data sets (this corresponds to each different line style on color machines).

-nl

Turn off drawing lines. When used with -m, -M, -p, or -P this can be used to produce scatter plots. When used with -bar, it can be used to produce standard bar graphs.

-o output filename

The name of the file to place the print output in. Currently defaults to /tmp/t.ps. See also the -print option.

-p

Marks each data point with a small marker (pixel sized). This is usually used with the -nl option for scatter plots.

-P

Similar to -p but marks each pixel with a large dot.

-print

Bring up the print dialog immediately upon startup. Unfortunately, there is no way to automatically print in JDK1.1, the user must hit the Ok button. See also the -o option.

-rv

Reverse video. On black and white displays, this will invert the foreground and background colors. The behaviour on color displays is undefined.

-t <string>

Title of the plot. This string is centered at the top of the graph.

-tf <fontname>

Title font. This is the name of the font to use for the graph title. See the -lf description above for how to specify fonts. The default is helvetica-BOLD-14

-tk

This option causes pxgraph to draw tick marks rather than full grid lines. The -bb option is also useful when viewing graphs with tick marks only.

-x <unitname>

This is the unit name for the X axis. Its default is "X".

-y <unitname>

This is the unit name for the Y axis. Its default is "Y".

-zg <color>

This is the color used to draw the zero grid line. Unsupported in the Java version.

-zw <width>

This is the width of the zero grid line in pixels. Unsupported in the Java version.

Compatibility Issues

Various compatibility issues are documented above in bold. Below are some other issues:

1. The original xgraph program allowed many formatting directives inside the file. This version only supports draw and move.
2. To read from standard input, specify a dash on the command line. Note that mixing reading from standard in and from files is not well supported.
3. This original xgraph program allowed blank lines to separate datasets. This version does not. Instead, use the move X Y directive.
4. This version does not support X resources.
5. The Java version of pxgraph takes longer to start up than the X11 version. This is an inherent problem with standalone Java applications. One guess is that most of the startup time comes from paging in the shared libraries.

For further information about this tool, see the Java Plot Website.

Since:

Ptolemy II 0.4

Version:

$Id: PxgraphParser.java 57040 2010-01-27 20:52:32Z cxh $

Author:

Edward A. Lee and Christopher Hylands

See Also:

PxgraphApplication, PxgraphApplet

Accepted Rating:

red (cxh)

Proposed Rating:

red (eal)

Field Summary

private static int	_BIG_ENDIAN
private boolean	_binary
private boolean	_connected
protected int	_currentdataset The current dataset, used for handling multiple files.
private int	_endian
private boolean	_firstInSet
private static int	_LITTLE_ENDIAN
private static int	_NATIVE_ENDIAN
protected Plot	_plot The plot object to which to apply commands.
private boolean	_sawFirstDataset

Constructor Summary

PxgraphParser(Plot plot)
Construct a parser to configure the specified plot.

Method Summary

private boolean	_addLegendIfNecessary(boolean connected)
private double[]	_parseDoubles(java.lang.String spec)
int	parseArgs(java.lang.String[] args) Parse pxgraph style command-line arguments.
int	parseArgs(java.lang.String[] args, java.net.URL base) Parse pxgraph style command-line arguments, using the specified base URL for any relative URL references.
int	parsePxgraphargs(java.lang.String pxgraphargs, java.net.URL base) Split a string containing pxgraph-compatible command-line arguments into an array and call parseArgs() on the array.
void	read(java.io.InputStream inputStream) Read a pxgraph-compatible binary or ASCII encoded file.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

_currentdataset

protected int _currentdataset

The current dataset, used for handling multiple files.

_plot

protected Plot _plot

The plot object to which to apply commands.

_NATIVE_ENDIAN

private static final int _NATIVE_ENDIAN

See Also:

Constant Field Values

_BIG_ENDIAN

private static final int _BIG_ENDIAN

See Also:

Constant Field Values

_LITTLE_ENDIAN

private static final int _LITTLE_ENDIAN

See Also:

Constant Field Values

_binary

private boolean _binary

_connected

private boolean _connected

_endian

private int _endian

_firstInSet

private boolean _firstInSet

_sawFirstDataset

private boolean _sawFirstDataset

Constructor Detail

PxgraphParser

public PxgraphParser(Plot plot)

Construct a parser to configure the specified plot.

Parameters:

plot - The Plot object that is configured.

Method Detail

parseArgs

public int parseArgs(java.lang.String[] args)
              throws CmdLineArgException,
                     java.io.FileNotFoundException,
                     java.io.IOException

Parse pxgraph style command-line arguments.

Parameters:

args - A set of command-line arguments.

Returns:

The number of arguments read.

Throws:

CmdLineArgException - If there is a problem parsing the command line arguments.

java.io.FileNotFoundException - If a file is specified that is not found.

java.io.IOException - If an error occurs reading an input file.

parseArgs

public int parseArgs(java.lang.String[] args,
                     java.net.URL base)
              throws CmdLineArgException,
                     java.io.FileNotFoundException,
                     java.io.IOException

Parse pxgraph style command-line arguments, using the specified base URL for any relative URL references.

Parameters:

args - A set of command-line arguments.

base - A base URL for relative URL references, or null if there is none.

Returns:

The number of arguments read.

Throws:

CmdLineArgException - If there is a problem parsing the command line arguments.

java.io.FileNotFoundException - If a file is specified that is not found.

java.io.IOException - If an error occurs reading an input file.

parsePxgraphargs

public int parsePxgraphargs(java.lang.String pxgraphargs,
                            java.net.URL base)
                     throws CmdLineArgException,
                            java.io.FileNotFoundException,
                            java.io.IOException

Split a string containing pxgraph-compatible command-line arguments into an array and call parseArgs() on the array. This is used in the rare circumstance that you want to control the format of a plot from an applet HTML file rather than in the plot data file.

Parameters:

pxgraphargs - The command line arguments.

base - A base URL for relative URL references, or null if there is none.

Returns:

The number of arguments read.

Throws:

CmdLineArgException - If there is a problem parsing the command line arguments.

java.io.FileNotFoundException - If a file is specified that is not found.

java.io.IOException - If an error occurs reading an input file.

read

public void read(java.io.InputStream inputStream)
          throws java.io.IOException

Read a pxgraph-compatible binary or ASCII encoded file.

Parameters:

inputStream - The input stream.

Throws:

java.io.IOException - If an I/O error occurs.

_addLegendIfNecessary

private boolean _addLegendIfNecessary(boolean connected)

_parseDoubles

private double[] _parseDoubles(java.lang.String spec)