use PDL::Graphics::Prima::Simple

To get started, you will want to use PDL::Graphics::Prima::Simple. This module provides a set of friendly wrappers for simple, first-cut data visualization. PDL::Graphics::Prima, the underlying library, is a general-purpose 2D plotting library built as a widget in the Prima GUI toolkit, but we don't need the full functionality for the purposes of this demo.

 use PDL::Graphics::Prima::Simple;
 my $x = sequence(100)/10;
 line_plot($x, $x->sin);

More than just lines!

In addition to numerous ways to plot x/y data, you can also plot distributions and images. The best run-down of the simple plotting routines can be found in the Synopsis for PDL::Graphics::Prima::Simple.

 $distribution = grandom(100);
 hist_plot($distribution);
 
 $x = sequence(100)/10;
 cross_plot($x, $x->sin);
 
 $image = rvals(100, 100);
 matrix_plot($image);

Mouse Interaction

Plots allow for mouse interaction, herein referred to as twiddling. You can resize the window, zoom with the scroll wheel, or click and drag the canvas around. There is also a right-click zoom-rectangle, and a right-click context menu.

 hist_plot(grandom(100));
 
 # Run this, then try using your mouse

In your Perl scripts, and in the PDL shell for some operating systems and some versions of Term::ReadLine, twiddling will cause your script to pause when you create a new plot. To resume your script or return execution to the shell, either close the window or press 'q'.

 # If your PDL shell supports simultaneous
 # input and plot interaction, running this
 # should display both plots simultaneously:
 
 $x = sequence(100)/10;
 cross_plot($x, $x->sin);
 line_plot($x, $x->cos);

Multiple plots without blocking

The blocking behavior just discussed is due to what is called autotwiddling. To turn this off, simply send a boolean false value to auto_twiddle. Then, be sure to invoke twiddling when you're done creating your plots.

 auto_twiddle(0);
 hist_plot(grandom(100));
 matrix_plot(rvals(100, 100));
 twiddle();

Once turned off, autotwiddling will remain off until you turn it back on.

 # autotwiddling still off
 hist_plot(grandom(100));
 matrix_plot(rvals(100, 100));
 twiddle();

Adding a title and axis labels

Functions like hist_plot, cross_plot, and matrix_plot actually create and return plot objects which you can subsequently modify. For example, adding a title and axis labels are pretty easy. For titles, you call the title method on the plot object. For axis labels, you call the label method on the axis objects.

 # Make sure autotwiddling is off in your script
 auto_twiddle(0);
 
 # Build the plot
 my $x = sequence(100)/10;
 my $plot = line_plot($x, $x->sin);
 
 # Add the title and labels
 $plot->title('Harmonic Oscillator');
 $plot->x->label('Time [s]');
 $plot->y->label('Displacement [cm]');
 
 # Manually twiddle once everything is finished
 twiddle();

Saving to a file

PDL::Graphics::Prima::Simple excels at user interaction, but you can save your plots to a file using save_to_file or save_to_postscript methods, or by right-clicking and selecting the appropriate menu option.

 auto_twiddle(0);
 $x = sequence(100)/10;
 line_plot($x, $x->sin)->save_to_postscript;
 
 # You can supply a filename to the method if you like.
 # Also available is save_to_file, which saves to raster
 # file formats. Expect save_to_postscript to be merged
 # into save_to_file in the future.

Adding additional data to the plot

Once you have created a plot, you can add additional data to it. You achieve this by adding a new DataSet with the data you want displayed.

 auto_twiddle(0);
 my $plot = hist_plot(grandom(100));
 
 # Add a Gaussian curve that "fits" the data
 use PDL::Constants qw(PI);
 my $fit_xs = zeroes(100)->xlinvals(-2, 2);
 my $fit_ys = exp(-$fit_xs**2 / 2) / sqrt(2*PI);
 $plot->dataSets->{fit_curve} = ds::Pair($fit_xs, $fit_ys);
 
 twiddle();

The default plot type for pairwise data is Diamonds. You can choose a different pairwise plot type, or even mix and match multiple pairwise plot types.

 auto_twiddle(0);
 my $plot = hist_plot(grandom(100));
 
 # Add a Gaussian curve that "fits" the data
 use PDL::Constants qw(PI);
 my $fit_xs = zeroes(200)->xlinvals(-5, 5);
 my $fit_ys = exp(-$fit_xs**2 / 2) / sqrt(2*PI);
 $plot->dataSets->{fit_curve} = ds::Pair($fit_xs, $fit_ys,
     # Use lines
     plotTypes => [
         ppair::Lines(
             # with a thickness of three pixels
             lineWidth => 3,
             # And the color red
             color => cl::LightRed,
         ),
         ppair::Diamonds,
     ],
 );
 
 twiddle();

The plot command

If you want to specify everything in one command, you can use the plot function. This lets you put everything together that we've already discussed, including multiple DataSets in a single command, title specification, and x and y axis options.

 # Generate some data:
 my $xs = sequence(100)/10 + 0.1;
 my $ys = $xs->sin + $xs->grandom / 10;
 my $y_err = $ys->grandom/10;
 
 # Plot the data and the fit
 plot(
     -data => ds::Pair($xs, $ys,
         plotTypes => [
             ppair::Triangles(filled => 1),
             ppair::ErrorBars(y_err => $y_err),
         ],
     ),
     -fit  => ds::Func(\&PDL::sin,
         lineWidth => 3,
         color => cl::LightRed,
     ),
     -note => ds::Note(
         pnote::Text('Incoming Signal',
             x => 0.2,
             y => sin(0.2) . '-3em',
         ),
     ),
     title => 'Noisey Sine Wave',
     x => {
         label => 'Time [s]',
         scaling => sc::Log,
     },
     y => { label => 'Measurement [Amp]' },
 );

Enjoy PDL::Graphics::Prima!

I hope you've enjoyed the tour, and I hope you find PDL::Graphics::Prima to be a useful plotting tool!

 # Thanks!