CHA PTE R 5 in Java

Creator QR Code JIS X 0510 in Java CHA PTE R 5

successful in CGI programs and other web applications, in which the client on the other end can likely display PNG, GIF and JPEG images.

if ($gd_object->can('gif')) { # save image as GIF open(IMAGE, '>image.gif') or die "open >image.gif: $!"; binmode IMAGE; print IMAGE $chart->gd->gif; close IMAGE; } else { # save image as PNG open(IMAGE, '>image.png') or die "open >image.png: $!"; binmode IMAGE; print IMAGE $chart->gd->png; close IMAGE; }

This can be awkward. Fortunately, GD::Graph provides the export_format() method. This method will return the string png if the underlying GD library exports PNG data, and the string gif if it exports GIF format images. This string can be used both to dynamically determine what the extension of the saved file should be, as well as the name of the method to use to get at that data. Perl s dynamic nature allows you to do the following:

my $format = $chart->export_format; open(IMAGE, ">image.$format") or die "open >image.$format: $!"; binmode IMAGE; print IMAGE $chart->gd->$format(); close IMAGE;

which is as short as it gets. The GIFgraph and Chart::PNGgraph modules are different: in the first place the plot method of both these modules does not return a GD object, but does return the image data directly. Secondly, both modules provide a convenient method to directly plot your image to a file. For GIFgraph:

my $img_data = $chart->plot(\@data); open(IMAGE, '>file.gif') or die $!; binmode IMAGE; print IMAGE $img_data; close IMAGE; # or $chart->plot_to_gif('file2.gif', \@data);

Chart::PNGgraph provides the same functionality, but of course for PNG, and the method is called plot_to_png(). GD::GRAPH AND FRIENDS 67

Bar charts are often used in financial reports and statistical comparisons. Most commonly the X axis displays the labels for each of the bars, such as month names or species names, and the distance between the bars has little significance. But only rarely do you see bar charts used in a situation in which the X axis has a numerical value that is not equidistantly spaced. GD::Graph can create bar charts for multiple data sets in one figure. If you have more than one set, you might choose to display the bars next to each other, in front of each other, or on top of each other. Displaying bars in front of each other is useful mainly to clearly offset a difference between one data set and another, when one is consistently smaller than the other. The visible part of the rear dataset shows the difference. Placing bars on top of each other can be used, for example in financial charts, where one wants to show the cumulative effect of multiple sources of income or expenditure. An example of a simple bar chart can be seen in figure 5.1. This chart was produced with the code:8

Figure 5.1 A simple bar chart, created with GD::Graph, illustrating the basic capabilities of the module.

use GD::Graph::bars; my @data = ( ['frogs','fish','toads','rodents'], [ 8, 3, 1, 2], ); my $chart = GD::Graph::bars->new(700,500); $chart->set(title => 'Ostrich diet'); $chart->set_title_font('arialbd', 24); $chart->set_x_axis_font('arial', 16); $chart->set_y_axis_font('arial', 16); my $gd_object = $chart->plot(\@data);

Note that the setting of the fonts is optional. It is done here purely for aesthetic considerations.

Sometimes you ll want to show the cumulative effect of sets of data. For example, you have a log file from a web server and you filter out all the unimportant requests, such as images and the like. You are left with only requests which are important to you, and want to create a chart that expresses the total number of requests in each hour of the day. You are also interested in how these requests are divided between GET and POST requests, and how these are further split between the secure and nonsecure parts of your site.

Figure 5.2 Two representations of one day of HTTP server log statistics, split up by the different types of requests. The data sets are cumulative, each band showing the respective contribution to the total of each type of request. (a)

Of course you can create a few bar charts and some pie charts that show you all of the above. However, it is possible to do all of this in one figure, if you treat the data cumulatively as shown in figure 5.2. Suppose you have already created summary data for the number of requests per hour of the day, divided them into the relevant groups, and note that this data lives in a file with tab-separated fields of the following format:

# hour GET POST 0 1495 1057 1 1106 680 2 1120 757 # and more lines... SSL_GET 955 819 455 SSL_POST 278 278 94