Data::Report - Framework for flexible reporting

  use Data::Report;

  # Create a new reporter.
  my $rep = Data::Report::->create(type => "text"); # or "html", or "csv", ...

  # Define the layout.
  $rep->set_layout
    ([ { name => "acct", title => "Acct",        width => 6  },
       { name => "desc", title => "Description", width => 40, align => "<" },
       { name => "deb",  title => "Debet",       width => 10, align => ">" },
       { name => "crd",  title => "Credit",      width => 10, align => ">" },
     ]);

  # Start the reporter.
  $rep->start;

  # Add data, row by row.
  $rep->add({ acct => 1234, desc => "Received", deb => "242.33"                  });
  $rep->add({ acct => 5678, desc => "Paid",                      crd => "699.45" });
  $rep->add({ acct => 1259, desc => "Taxes",    deb =>  "12.00", crd => "244.00" });
  $rep->add({               desc => "TOTAL",    deb => "254.33", crd => "943.45" });

  # Finish the reporter.
  $rep->finish;

Data::Report is a flexible, plugin-driven reporting framework. It makes it easy to define reports that can be produced in text, HTML and CSV. Textual ornaments like extra empty lines, dashed lines, and cell lines can be added in a way similar to HTML style sheets.

The Data::Report framework consists of three parts:

The plugins

Plugins implement a specific type of report. Standard plugins provided are Data::Report::Plugin::Text for textual reports, Data::Report::Plugin::Html for HTML reports, and Data::Report::Plugin::Csv for CSV (comma-separated) files.

Users can, and are encouraged, to develop their own plugins to handle different styles and types of reports.

The base class

The base class Data::Report::Base implements the functionality common to all reporters, plus a number of utility functions the plugins can use.

The factory

The actual Data::Report module is a factory that creates a reporter for a given report type by selecting the appropriate plugin and returning an instance thereof.

Note that except for the create method, all other methods are actually handled by the plugins and their base class.

Reporter objects are created using the class method create. This method takes a hash (or hashref) of arguments to initialise the reporter object.

The actual reporter object is implemented by one of the plugin modules, selected by the type argument. Standard plugins are provided for text, HTML and CSV report types. The default type is text.

When looking for a plugin to support report type foo, the create method will first try to load a module My::Package::Foo where My::Package is the invocant class. If this module cannot be loaded, it will fall back to Data::Report::Plugin::Foo. Note that, unless subclassed, the current class will be Data::Report.

All other initialisation arguments correspond to attribute setting methods provided by the plugins. For example, the hypothetical call

  my $rpt = Data::Report->create(foo => 1, bar => "Hello!");

is identical to:

  my $rpt = Data::Report->create;
  $rpt->set_foo(1);
  $rpt->set_bar("Hello!");

You can choose any combination at your convenience.

This method indicates that all setup has been completed, and starts the reporter. Note that no output is generated until the add method is called.

start takes no arguments.

Although this method could be eliminated by automatically starting the reporter upon the first call to add, it turns out that an aplicit start makes the API much cleaner and makes it easier to catch mistakes.

This method adds a new entry to the report. It takes one single argument, a hash ref of column names and the corresponding values. Missing columns are left blank.

In addition to the column names and values, you can add the special key _style to designate a particular style for this entry. What that means depends on the plugin that implements this reporter. For example, the standard HTML reporter plugin prefixes the given style with r_ to form the class name for the row. The style name should be a simple name, containing letters, digits and underscores, starting with a letter.

Example

  $rpt->add({ date   => "2006-04-31",
              amount => 1000,
              descr  => "First payment",
              _style => "plain" });

This method indicates that report generation is complete. After this, you can call start again to initiate a new report.

finish takes no arguments.

This is a convenience method. If the output stream was set up by the reporter itself (see set_output, below), the stream will be closed. Otherwise, this method will be a no-op.

close takes no arguments.

The reporter type.

This is the most important attribute, since it effectively defines the report layout.

This method takes one argument, an array reference. Each element of the array is a hash reference that corresponds to one column in the report. The order of elements definines the order of the columns in the report, but see set_fields below.

The following keys are possible in the hash reference:

name

The name of this column. The name should be a simple name, containing letters, digits and underscores, starting with a letter.

The standard HTML reporter plugin uses the column name to form a class name for each cell by prefixing with c_. Likewise, the classes for the table headings will be formed by prefixing the column names with h_. See ADVANCED EXAMPLES, below.

title

The title of this column. This title is placed in the column heading.

width

The width of this column. Relevant for textual reporters only.

By default, if a value does not fit in the given width, it will be spread over multiple rows in a pseudo-elegant way. See also the truncate key, below.

align

The alignment of this column. This can be either < for left-aligned columns, or > to indicate a right-aligned column.

truncate

If true, the values in this column will be truncated to fit the width of the column. Relevant for textual reporters only.

This method can be used to set an arbitrary style (a string) whose meaning depends on the implementing plugin. For example, a HTML plugin could use this as the name of the style sheet to use.

The name should be a simple name, containing letters, digits and underscores, starting with a letter.

Returns the style, or default if none.

Designates the destination for the report. The argument can be

a SCALAR reference

All output will be appended to the designated scalar.

an ARRAY reference

All output lines will be pushed onto the array.

a SCALAR

A file will be created with the given name, and all output will be written to this file. To close the file, use the close method described above.

anything else

Anything else will be considered to be a file handle, and treated as such.

The stylist is a powerful method to control the appearance of the report at the row and cell level. The basic idea is taken from HTML style sheets. By using a stylist, it is possible to add extra spaces and lines to rows and cells in a declarative way.

When used, the stylist should be a reference to a possibly anonymous subroutine with three arguments: the reporter object, the style of a row (as specified with _style in the add method), and the name of a column as defined in the layout. For table headings, the row name _head is used.

The stylist routine will be repeatedly called by the reporter to obtain formatting properties for rows and cells. It should return either nothing, or a hash reference with properties.

When called with only the row argument, it should return the properties for this row.

When called with row equal to "*" and a column name, it should return the properties for the given column.

When called with a row and a column name, it should return the properties for the given row/column (cell).

All appropriate properties are merged to form the final set of properties to apply.

The following row properties are recognised. Between parentheses the backends that support them.

skip_before

(Text) Produce an empty line before printing the current row.

skip_after

(Text) Produce an empty line after printing the current row, but only if other data follows.

line_before

(Text) Draw a line of dashes before printing the current row.

line_after

(Text) Draw a line of dashes after printing the current row.

cancel_skip

(Text) Cancel the effect of a pending skip_after

ignore

(All) Ignore this row. Useful for CSV backends where only the raw data matters, and not the totals and such.

The following cell properties are recognised. Between parentheses the backends that support them.

indent

(Text) Indent the contents of this cell with the given amount.

wrap_indent

(Text) Indent wrapped contents of this cell with the given amount.

truncate

(Text) If true, truncate the contents of this cell to fit the column width.

line_before

(Text) Draw a line in the cell before printing the current row. The value of this property indicates the symbol to use to draw the line. If it is 1, dashes are used.

line_after

(Text) Draw a line in the cell after printing the current row. The value of this property indicates the symbol to use to draw the line. If it is 1, dashes are used.

raw_html

(Html) Do not escape special HTML characters, allowing pre-prepared HTML code to be placed in the output. Use with care.

ignore

(All) Ignore this column. Note that to prevent surprising results, the column must be ignored in all applicable styles, including the special style "_head" that controls the heading.

class

(Html) Class name to be used for this cell. Default class name is "h_CNAME" for table headings and "c_CNAME" for table rows, where CNAME is the name of the column.

Example:

  $rep->set_stylist(sub {
    my ($rep, $row, $col) = @_;

    unless ( $col ) {
	return { line_after => 1 } if $row eq "total";
	return;
    }
    return { line_after => 1 } if $col eq "amount";
    return;
  });

Each reporter provides a standard (dummy) stylist called _std_stylist. Overriding this method is equivalent to using set_stylist.

Returns the current stylist, if any.

Headings consist of two parts, the top heading, and the standard heading. Bij default, the top heading is empty, and the standard heading has the names of the columns with a separator line (depnendent on the plugin used).

This method can be used to designate a subroutine that will provide the top heading of the report.

Example:

  $rpt->set_topheading(sub {
    my $self = shift;
    $self->_print("Title line 1\n");
    $self->_print("Title line 2\n");
    $self->_print("\n");
  });

Note the use of the reporter provided _print method to produce output.

When subclassing a reporter, a method _top_heading can be defined to provide the top heading. This is equivalent to an explicit call to set_topheading, but doesn't need to be repeatedly and explicitly executed for each new reporter.

Returns the current top heading routine, if any.

This method can be used to designate a subroutine that provides the standard heading of the report.

In normal cases using this method is not necessary, since setting the top heading will be sufficient.

Each reporter plugin provides a standard heading, implemented in a method called _std_header. This is the default value for the heading attribute. A user-defined heading can use

  $self->SUPER::_std_header;

to still get the original standard heading produced.

Example:

  $rpt->set_heading(sub {
    my $self = shift;
    $self->_print("Title line 1\n");
    $self->_print("Title line 2\n");
    $self->_print("\n");
    $self->SUPER::_std_heading;
    $self->_print("\n");
  });

Note the use of the reporter provided _print method to produce output.

When subclassing a reporter, the method _std_heading can be overridden to provide a customized top heading. This is equivalent to an explicit call to set_topheading, but doesn't need to be repeatedly and explicitly executed for each new reporter.

Returns the current standard heading routine, if any.

This method can be used to define what columns (fields) should be included in the report and the order they should appear. It takes an array reference with the names of the desired columns.

Example:

  $rpt->set_fields([qw(descr amount date)]);

Returns the current set of selected columns.

This method defines the width for one or more columns. It takes a hash reference with column names and widths. The width may be an absolute number, a relative number (to increase/decrease the width, or a percentage.

Example:

  $rpt->set_width({ amount => 10, desc => '80%' });

Returns a hash with all column names and widths.

This example subclasses Data::Report with an associated plugin for type text. Note the use of overriding _top_heading and _std_stylist to provide special defaults for this reporter.

  package POC::Report;

  use base qw(Data::Report);

  package POC::Report::Text;

  use base qw(Data::Report::Plugin::Text);

  sub _top_heading {
      my $self = shift;
      $self->_print("Title line 1\n");
      $self->_print("Title line 2\n");
      $self->_print("\n");
  }

  sub _std_stylist {
      my ($rep, $row, $col) = @_;

      if ( $col ) {
	  return { line_after => "=" }
	    if $row eq "special" && $col =~ /^(deb|crd)$/;
      }
      else {
	  return { line_after => 1 } if $row eq "total";
      }
      return;
  }

It can be used as follows:

  my $rep = POC::Report::->create(type => "text");

  $rep->set_layout
    ([ { name => "acct", title => "Acct",   width => 6  },
       { name => "desc", title => "Report", width => 40, align => "<" },
       { name => "deb",  title => "Debet",  width => 10, align => "<" },
       { name => "crd",  title => "Credit", width => 10, align => ">" },
     ]);

  $rep->start;

  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "normal" });
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "normal" });
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "special"});
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "total"  });

  $rep->finish;

The output will look like:

  Title line 1
  Title line 2

  Acct                                      Report  Debet           Credit
  ------------------------------------------------------------------------
  one                                          two  three             four
  one                                          two  three             four
  one                                          two  three             four
                                                    ==========  ==========
  one                                          two  three             four
  ------------------------------------------------------------------------

This is a similar example for a HTML reporter:

  package POC::Report;

  use base qw(Data::Report);

  package POC::Report::Html;

  use base qw(Data::Report::Plugin::Html);

  sub start {
      my $self = shift;
      $self->{_title1} = shift;
      $self->{_title2} = shift;
      $self->{_title3} = shift;
      $self->SUPER::start;
  }

  sub _top_heading {
      my $self = shift;
      $self->_print("<html>\n",
		    "<head>\n",
		    "<title>", $self->_html($self->{_title1}), "</title>\n",
		    '<link rel="stylesheet" href="css/', $self->get_style, '.css">', "\n",
		    "</head>\n",
		    "<body>\n",
		    "<p class=\"title\">", $self->_html($self->{_title1}), "</p>\n",
		    "<p class=\"subtitle\">", $self->_html($self->{_title2}), "<br>\n",
		    $self->_html($self->{_title3}), "</p>\n");
  }

  sub finish {
      my $self = shift;
      $self->SUPER::finish;
      $self->_print("</body>\n</html>\n");
  }

Note that it defines an alternative start method, that is used to pass in additional parameters for title fields.

The method _html is a convenience method provided by the framework. It returns its argument with sensitive characters escaped by HTML entities.

It can be used as follows:

  package main;

  my $rep = POC::Report::->create(type => "html");

  $rep->set_layout
    ([ { name => "acct", title => "Acct",   width => 6  },
       { name => "desc", title => "Report", width => 40, align => "<" },
       { name => "deb",  title => "Debet",  width => 10, align => "<" },
       { name => "crd",  title => "Credit", width => 10, align => ">" },
     ]);

  $rep->start(qw(Title_One Title_Two Title_Three_Left&Right));

  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "normal" });
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "normal" });
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "normal" });
  $rep->add({ acct => "one", desc => "two", deb => "three", crd => "four", _style => "total"  });

  $rep->finish;

The output will look like this:

  <html>
  <head>
  <title>Title_One</title>
  <link rel="stylesheet" href="css/default.css">
  </head>
  <body>
  <p class="title">Title_One</p>
  <p class="subtitle">Title_Two<br>
  Title_Three_Left&amp;Right</p>
  <table class="main">
  <tr class="head">
  <th align="left" class="h_acct">Acct</th>
  <th align="left" class="h_desc">Report</th>
  <th align="right" class="h_deb">Debet</th>
  <th align="right" class="h_crd">Credit</th>
  </tr>
  <tr class="r_normal">
  <td align="left" class="c_acct">one</td>
  <td align="left" class="c_desc">two</td>
  <td align="right" class="c_deb">three</td>
  <td align="right" class="c_crd">four</td>
  </tr>
  <tr class="r_normal">
  <td align="left" class="c_acct">one</td>
  <td align="left" class="c_desc">two</td>
  <td align="right" class="c_deb">three</td>
  <td align="right" class="c_crd">four</td>
  </tr>
  <tr class="r_normal">
  <td align="left" class="c_acct">one</td>
  <td align="left" class="c_desc">two</td>
  <td align="right" class="c_deb">three</td>
  <td align="right" class="c_crd">four</td>
  </tr>
  <tr class="r_total">
  <td align="left" class="c_acct">one</td>
  <td align="left" class="c_desc">two</td>
  <td align="right" class="c_deb">three</td>
  <td align="right" class="c_crd">four</td>
  </tr>
  </table>
  </body>
  </html>

See also the examples in t/09poc*.t.

Johan Vromans, <jvromans at squirrel.nl>

Disclaimer: This module is derived from actual working code, that I turned into a generic CPAN module. During the process, some features may have become unstable, but that will be cured in time. Also, it is possible that revisions of the API will be necessary when new functionality is added.

Please report any bugs or feature requests to bug-data-report at rt.cpan.org, or through the web interface at http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Data-Report. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

You can find documentation for this module with the perldoc command.

    perldoc Data::Report          (user API)
    perldoc Data::Report::Base    (plugin writer documentation)

You can also look for information at:

* RT: CPAN's request tracker

http://rt.cpan.org/NoAuth/Bugs.html?Dist=Data-Report

* CPAN Ratings

http://cpanratings.perl.org/d/Data-Report

* Search CPAN

http://search.cpan.org/dist/Data-Report

Copyright 2006,2008 Squirrel Consultancy, all rights reserved.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.