=pod

=head1 NAME

PDF::Table - A utility class for building table layouts in a PDF::Builder 
(or PDF::API2) object.

=head1 SYNOPSIS

Rather than cluttering up the following documentation with B<(or PDF::API2)>
additions, wherever it refers to C<PDF::Builder>, understand that you can 
substitute C<PDF::API2> to use that product instead I<EXCEPT where noted>.

 use PDF::Builder;
 use PDF::Table;

 my $pdftable = new PDF::Table;
 my $pdf = new PDF::Builder(-file => "table_of_lorem.pdf");
 my $page = $pdf->page();

 # some data to lay out
 my $some_data =[
    ["1 Lorem ipsum dolor",
    "Donec odio neque, faucibus vel",
    "consequat quis, tincidunt vel, felis."],
    ["Nulla euismod sem eget neque.",
    "Donec odio neque",
    "Sed eu velit."],
    # ... and so on
 ];

 $left_edge_of_table = 50;
 # build the table layout
 # note: ignoring return values array
 $pdftable->table(
     # required parameters
     $pdf,
     $page,
     $some_data,
     'x' => $left_edge_of_table,
     'w' => 495,
     'y' => 500,
     'h' => 300,
     # some optional parameters
     'next_y'          => 750,
     'next_h'          => 500,
     'padding'         => 5,
     'padding_right'   => 10,
     'bg_color_odd'    => "gray",
     'bg_color_even'   => "lightblue", # cell bg color for even rows
     'max_word_length' => 50, # 50 between forced splits
  );

 # do other stuff with $pdf
 $pdf->save();
...

=head2 EXAMPLE

For a complete working example or initial script look into distribution's 
'examples' folder.

=head1 DESCRIPTION

This class is a utility for use with the PDF::Builder (or PDF::API2, see
note above) module from CPAN.
It can be used to display text data in a table layout within a PDF.
The text data must be in a 2D array (such as returned by a DBI statement 
handle C<fetchall_arrayref()> call).
PDF::Table will automatically add as many new pages as necessary to display 
all of the data.
Various layout properties, such as font, font size, cell padding, and 
background color can be specified for each column and/or for even/odd rows.
Also a (non)repeated header row with different layout properties can be 
specified.

See the L</METHODS> section for complete documentation of every parameter.

=head1 COMPATIBILITY

Starting with version 1.000, several behaviors have changed (for the better, I
believe). Nevertheless, there may be some users who prefer the old behaviors.
To keep everybody happy, it is possible to easily revert to the old behaviors.
Near the top of Table.pm, look for a section labeled C<COMPATIBILITY WITH OLDER 
VERSIONS>. You can change settings here to match old behaviors:

=over

=item repeating headers

The old default for the C<repeat> setting for a header was '0' (do not repeat
after a table has been split across a page). I believe that most users will
want to automatically repeat a header row at the start of each table fragment, 
but you can change this behavior if you wish. Change C<$repeat_default> from 1
to 0 to get the old behavior (or, explicitly give C<repeat =E<gt> 0> in the 
header properties settings).

=item which rows are 'odd' (and which are 'even')

PDF::Table decided which rows were odd/even (background and foreground colors,
etc) in an inconsistent manner, especially if a header was used (whether 
repeated or not). Now, the first data row (excluding headers) is "odd", and
all rows after that alternate "even", "odd", etc., even across page breaks. If 
you want the old behavior, it can be requested. Change C<$oddeven_default> from
1 to 0 to get the old behavior.

=item default cell padding

The old default for padding around the contents of a cell was 0. It is now
2pt. Change C<$padding_default> from 2 to 0 to get the old behavior.

=item behavior of borders

The old behavior was calling both the frame around the table I<and> the
cell-divider rules as "border", and using the same settings for both. This has
been changed to separate the two classes, with "border" referring to the outside
framework, and "rules" referring to the dividers. Note that "rules" still
inherit from "border", so an explicit definition of C<rules =E<gt> 0> (to hide
interior rules) or another width (line weight) may still be needed to override 
the "border" setting for interior dividers.

=back

=head2 Maintaining compatibility

Near the top of file Table.pm, look for C<my $compat_mode = 0;>.
PDF::Table is shipped with a flag of C<0> to use the new features of the 
library. If you have a pressing need to maintain compatibility with older
versions of the library, you may change the value to C<1>.
Note that a flag of C<1> will break some of the t-tests, because of different
padding defaults resulting in different text locations on the page.

=head2 Run-time changes

If you do not wish to change the PDF::Table code section to permanently change
old-versus-new behavior, you can use the I<compatibility> flag in the settings
to temporarily change the variables listed above.

    compatibility => [ 0, 0, 0 ]

will restore all behaviors to the old style, while 

    compatibility => [ 1, 0, 2 ]

will change only the designation of "odd/even" rows (element 1) to the old 
behavior, while leaving header repeat (element 0) and default padding (element
2) in the new behavior.

=head1 METHODS

=head2 new()

    my $pdf_table = new PDF::Table;
       or
    my $pdf_table = PDF::Table->new();

=over

=item Description

Creates a new instance of the class.

=item Parameters

There are no required parameters. You may pass $pdf, $page, $data, and
%options; or can defer this until the table() method invocation (the usual
technique).

=item Returns

Reference to the new instance

=back

=head2 table()

Generate output:

    my ($final_page, $number_of_pages, $final_y) = table($pdf, $page, $data, %settings)

or, just report row sizes:

    my @v_sizes = table($pdf, $page, $data, %settings, 'ink' => 0)

=over

=item Description

Generates a multi-row, multi-column table into an existing PDF document, based 
on provided data set and settings.

=item Parameters

    $pdf      - a PDF::Builder instance representing the document being created
    $page     - a PDF::Builder::Page instance representing the current page of 
                the document
    $data     - an ARRAY reference to a 2D data structure that will be used 
                to build the table
    %settings - HASH with geometry and formatting parameters

B<$data> is normally an m x n matrix of strings, each string containing the
content text of that cell, and arranged with each row being an anonymous
array of strings. For example, a table of 3 rows of 4 columns would be
 
    $data = [
      [ 'c1', 'c2', 'c3', 'c4' ],  # row 1
      [ 'c1', 'c2', 'c3', 'c4' ],  # row 2
      [ 'c1', 'c2', 'c3', 'c4' ],  # row 3
            ];

B<For PDF::Builder ONLY:>

If you wish to enable I<markup> (Markdown or HTML) for a cell, C<$data> has
the same structure, except that a cell with markup is itself an anonymous
array of three items: a string describing the markup ('none', 'md1', 'html',
or 'pre'), a string or anonymous array of strings representing the text and
markup for the cell, and an anonymous hash of settings pertaining to the
C<column()> call. See PDF::Builder's POD for Content::Text for details.

Instead of plain text cell 'This is plain text', you can do something like:

    [ 'html', 
      '<p><font color="green">This is green text.</font></p>',
      { 'font-family'=>'Serif', 'font-size'=>15, 'para'=>[0,0] } ]

Remember, this works only with PDF::Builder, and I<not> with PDF::API2.
See the "Column.pl" example on PDF::Builder for some examples.
Automatic cell sizing will not work with this, so you need to use the
C<size> table parameter to specify relative cell widths.

B<End of PDF::Builder ONLY:>

For full C<%settings> description see section L</Table settings> below.

This method will add more pages to the PDF instance as required, based on the 
formatting options and the amount of data.

=item Returns

If actually creating PDF output, the return value is a 3 item list where

    $final_page - A PDF::Builder::Page instance that the table ends on
    $number_of_pages - The count of pages that the table spans
    $final_y - The Y coordinate of the table bottom, so that additional 
               content can be added on the same page ($final_page)

If just requesting the table size (C<'ink' =E<gt> 0>), the return value is an C<N+3>
item list where

    item [0] is the overall table length (height) in points
    item [1] is the header height in points (0 if no header used)
    item [2] is the footer height in points (currently 0)
    item [3] is the height of data row 1
    item [4] is the height of data row 2
     ...
    item [N+3] is the height of data row N

See C<examples/vsizes.pl> for some examples of determining sizes of a table and
its rows. This may be useful information for you in deciding whether there is
space on I<this> page to fit the complete table, or if you should go to the
top of the next page. It could be used to figure how to set I<h> to fit
complete rows on a page (not splitting rows), or even how to re-order the rows
to most efficiently display them (on the fewest number of pages) without 
splitting any row.

Two things to note:

1. The table data must be refreshed before calling the table method a second time (to actually output PDF, with ink = 1).

2. If the table spills over onto a new page when outputting the PDF (ink = 1), you will need to retrieve the new C<$page> hash via the return code. Actually, you would need to do this even if you are not querying table sizes!

=item Example

    my $pdf  = new PDF::Builder;
    my $page = $pdf->page();
    my $data = [
        ['foo1','bar1','baz1'],
        ['foo2','bar2','baz2']
    ];
    my %settings = (
        'x' => 10,
        'w' => 570,
        'y' => 220,
        'h' => 180,
    );

    my ($final_page, $number_of_pages, $final_y) = 
        $pdftable->table( $pdf, $page, $data, %options );

=back

=head3 Table settings

Unless otherwise specified, all dimensional and geometry units used are 
measured in I<points>. Line counts are not used anywhere.

"Even" rows start with the first data (non-header) row. Think of this first
row as number zero (an I<even> number). Even rows alternate with odd rows.
The odd/even flag is B<not> reset when a table is split across pages. If a
table fragment ends on an odd row, the next fragment (on the next page),
starting the next row, will
start with an even row. If a I<row> is split across pages, it will resume with
the same odd/even setting as on the previous page. If you desire to have the
old (previous) odd/even behavior, see L</COMPATIBILITY>.

The name (key) of any table setting hash element may be given with or
without a leading dash (hyphen). A leading dash is allowed for compatibility
with older versions of PDF::Table, but is B<DEPRECATED!> It is recommended 
that the dash be omitted in new code, and removed from old code before
November 2022.

B<Note:> if you use a deprecated setting name, or a setting beginning with a 
hyphen '-', PDF::Table will update the settings list with the preferred name.
It does this by inserting the item using the preferred, non-hyphen name, and 
then deletes the deprecated one. Due to peculiarities in the way Perl copies
arrays, hashes, and references; it is possible that your input settings hash
may end up being modified! This normally will not be a cause for concern, but
you should be aware of this behavior in case you wish to reuse all or part of
a PDF::Table settings list (hash) for other purposes (or another table) -- 
it may have been slightly modified.

Note that any "Color specifier" is not limited to a name (e.g., 'black') or
a 6-digit hex specification (e.g., '#3366CC'). See the PDF::Builder 
writeup on specifying colors for CMYK, L*a*b, HSV, and other methods.

=head4 Mandatory global settings

There are some mandatory parameters for setting table geometry and position 
on the first (initial) or only page of the table. It is up to you to tell
PDF::Table where to start (upper left corner) the table, and its width and
maximum height on this page.

=over

=item B<x> - X coordinate of upper left corner of the table. 

The left edge of the sheet (media) is 0. 
B<Note> that this C<X> will be used for any spillover of the table to 
additional page(s), so you cannot have spillover (continuation) rows 
starting at a different C<X>.

B<Value:> can be any number satisfying C<0 E<le> X < PageWidth>

B<Default:> No default value

    'x' => 10,

=item B<y> - Y coordinate of upper left corner of the table on the 
initial page.

B<Value:> can be any number satisfying C<0 < y < PageHeight> 
(depending on space availability when embedding a table)

B<Default:> No default value

    'y' => 327,

B<Deprecated name:> I<start_y> (will go away in the future!)

=item B<w> - width of the table starting from C<x>.

B<Note> that this C<width> will be used for any spillover of the table to 
additional page(s), so you cannot have spillover (continuation) rows with a 
different C<width>.

B<Value:> can be any number satisfying C<0 < w < PageWidth - x>

B<Default:> No default value

    'w'  => 570,

B<NOTE:> If PDF::Table finds that the table width needs to be increased to
accommodate the requested text and formatting, it will output a warning. This
could lead to undesired results. Possible solutions to keep the table from
being widened include:

    1) Increase table width (w)
    2) Decrease font size (font_size)
    3) Choose a narrower font
    4) Decrease "max_word_length" parameter, so long words are split into
        shorter chunks
    5) Rotate media to landscape (if it is portrait)
    6) Use a larger (wider) media size

=item B<h> - Height of the table on the initial (current) page.

Think of this as the I<maximum height> (Y dimension) of the start of the
table on this page. This would be the current C<Y> location less any bottom
margin. Normally you would let as much as possible fit on the page,
but it's possible that you might want to split the table at an earlier point,
to put more on the next (spill) page.

B<Value:> can be any number satisfying C<0 < h < PageHeight - Current Y position>

B<Default:> No default value

    'h' => 250,

B<Deprecated name:> I<start_h> (will go away in the future!)

=back

=head4 Optional settings

These are settings which are not absolutely necessary, although their use may
result in a much more pleasing appearance for the table. They all have a
"reasonable" default (or inheritance from another setting).

=head4 Optional Global Settings

These settings apply only to the entire table, and cannot be used to specify
cell, column, or row properties. A global setting may only occur once.

=over

=item B<next_h> - Height of the table on any additional page.

Think of this as the I<maximum height> (Y dimension) of any overflow 
(spill) table portions on following pages.
I<It is highly recommended that you
explicitly specify this setting as the full (body content) height of a page,
rather than having PDF::Table try to figure out a good value and B<give a
warning>.>

B<Value:> can be any number satisfying C<0 < next_h < PageHeight - y>

You need to leave a non-negative amount of space at the bottom of the page.

B<Default:> Media height * 80% (80% of the paper height)
You will receive a warning if C<next_h> is needed for a spill page and you
did not provide it!

    'next_h'  => 700,

=item B<next_y> - Y coordinate of upper left corner of the table at any 
additional page.

Think of this as the starting C<Y> position of any overflow 
(spill or continuation) table portions on following pages.
I<It is highly recommended that you
explicitly specify this setting to be at the top of the body content of a page,
rather than having PDF::Table try to figure out a good value and B<give a
warning>.>

B<Value:> can be any number satisfying C<0 < next_y < PageHeight>

B<Default:> Media height * 90% (10% down from the top of the paper)
You will receive a warning if C<next_y> is needed for a spill page and you
did not provide it!

    'next_y'  => 750,

=item B<ink> - Whether to actually write the table, or to measure the table size

You may want to know how high (tall) the actual table will be, before making
a decision whether to start the table on I<this> page, or another (typically,
the top of the I<next> page). You might want to know if there is space for two
or more small tables on this page, or where best to split a table (give an 
C<h> value to avoid splitting a row). You may also decide to rearrange data 
rows, if this will make better use of available space.

The default I<ink> value is I<1> (to output to the PDF file). If set to I<0>, 
only the height measurements (total, header, footer, each data row), in points,
will be returned.

B<Notes:>

1. No "footer" can currently be defined. Space is reserved in the output list for a possible future implementation.

2. No matter if the header is repeated on each page, only one instance will be returned in the heights list. Think of the table as fitting on a single page, no matter how long it is, so don't forget to add in the (repeated) header height for each expected page of output! 

3. If you have exceptionally heavy (thick) top and bottom border lines, you might want to add in some extra height (half the thickness of each) to the overall height of the table, to account for this.

B<Value:> 0 (return height only) or 1 (actually write the table)

B<Default:> 1 (write out the table)

    'ink'  => 0,

B<Caution:> Currently, C<ink> is ignored for cells with I<markup>; their content
will be written out regardless of C<ink>. We I<do> plan to support C<ink> for
markup cells in the not-too-distant future.

=item B<size> - Specified column widths.

The default behavior of PDF::Table is to calculate the width needed for each
column, based on the longest word, content size, and optional minimum and 
maximum settings. As an alternative, you may give a string specifying the
widths of all columns, one entry per column. An entry may be a fixed width
(number and unit), or a relative width to be allocated from remaining space
after the fixed width columns are removed from the table width.

If no unit is given, 'pt' (big points, 1/72 inch) are assumed. If no number is
given, '1' is assumed (not recommended, except for '*'). The following units
are supported:

=over

=item B<pt> - points (1/72 inch)

=item B<in> - inches (72 points)

=item B<cm> - centimeters (28.3 points)

=item B<mm> - millimeters (2.83 points)

=item B<em> - ems (width of 'M', depends on table default font)

=item B<ex> - exs (width of 'x', depends on table default font)

=back

These are all absolute (fixed) sizes. If using font-dependent units (ems and 
exs), be aware of the overall font and font-size in use by PDF::Table (default
is normally 12pt Times-Roman).

There are also relative or allocatable sizes, specified with the unit '*'.
After all fixed-size columns have been processed, whatever table width that is
left over will be distributed among the remaining (relative) columns 
proportional to the number for each column (e.g., B<2*> gets twice the width 
that B<*> gets).

B<Value:> a string containing one entry per column

Each entry is separated by one or more whitespaces (e.g., blanks). Numbers 
are unsigned (positive) and may be integers or real numbers (with decimal
point). The unit follows immediately (no spaces between number and unit).
Note that any column I<minimum> or I<maximum> width settings will be ignored!

B<Default:> None. If not given, widths will be calculated for the columns.

    'size'  => '* 3.5in 1.7* 4cm',

=item B<new_page_func> - CODE reference to a function that returns a 
PDF::Builder::Page instance. See section L<New Page Function Hook> below.

    'new_page_func'  => $code_ref,

=item B<cell_render_hook> - CODE reference to a function called with the 
current cell coordinates. See section L<Cell Render Hook> below.

    'cell_render_hook'  => $code_ref,

=item B<header_props> - HASH reference to specific settings for the Header row 
of the table. See section L</Header Row Properties> below.

    'header_props' => $hdr_props,

=item B<row_props> - HASH reference to specific settings for each row of 
the table. See section L</Row Properties> below.

    'row_props' => $my_row_props,

=item B<column_props> - HASH reference to specific settings for each column of 
the table. See section L</Column Properties> below.

    'column_props' => $col_props,

=item B<cell_props> - HASH reference to specific settings for each column of 
the table. See section L</Cell Properties> below.

    'cell_props' => $cell_props,

=item B<border_w> - Width of table border lines.

=item B<h_border_w> - Width of horizontal border lines (top and bottom of the
table). Overrides 'border_w' value for horizontal usage. Note that if the
table spills over onto following pages, only the very first top and very last 
bottom table border will be full width. Dividers on row boundaries will be
1pt wide ($border_w_default) solid lines, and where a row is divided within 
its content, a dashed (pattern $dashed_rule_default) 1pt wide line is used.

=item B<v_border_w> -  Width of vertical border lines. Overrides 
'border_w' value for vertical usage.

B<Value:> can be any positive number. When set to 0, it will disable 
border lines. This is the line thickness for drawing a border.

B<Default:> C<1>  ($border_w_default)

The I<border> is the B<outside> frame around the table. It does not enter into
table height or width calculations, so be sure to set your C<x> and C<w>
settings to allow for the width of vertical borders, and your C<y> or C<next_y>
and C<h> or C<next_h> settings to allow for the width (thickness or height) of 
the horizontal borders, especially if you make them more than a Point or two 
in thickness (line width).

    'border_w'     => 3,     # border width is 3
    'h_border_w'   => 1,     # horizontal borders will be 1, overriding 3
    'v_border_w'   => undef, # vertical borders will be 3, as it will 
                             # fall back to 'border_w'

Note that both borders and rules overlay the exact boundary between two cells
(i.e., the centerline). That is, one half of a rule or border will overlay the
adjoining cells. Rules do not expand the size of the table, although
borders will (by a total of their thickness/width). If you set particularly
thick (wide) rules, pay attention to adding some padding on the appropriate
side(s), so that valuable content is not overlaid. For cells along the outer
border, one half the width of a border will overlay the cell, so account for
this in the padding specification.

B<Deprecated names:> I<border> (now 'border_w'), 
I<horizontal_borders> (now 'h_border_w'), 
and I<vertical_borders> (now 'v_border_w'); will go away in the future!

=item B<border_c> -  Border color for all borders.

B<Value:> Color specifier as 'name' or '#rrggbb'

B<Default:> C<'black'> ($fg_color_default)

    'border_c' => 'red',

B<Deprecated name:> I<border_color> (will go away in the future!)

The same color is used for both the horizontal and vertical borders.

=back

=head4 Optional Cell, Column, Row, or Global Settings

These settings can be specified to apply to the entire table, or more
narrowly applied to the header row (in header_props hash), one or more rows
(in row_props array), one or more columns (in column_props array), or one
or more individual cells (in cell_props hash).

If a setting is specified in more than one place, the order of precedence is
as follows: a header property (header row only), followed by a cell property, 
followed by a column property, followed by a row property, followed by a 
global setting, and finally, any hard-coded default value (if required).

A global setting may only occur once (although it may be overridden by cell,
column, or row usage of the same setting).

=over

=item B<default_text> - A string to use if no content (text) is defined for
a cell.

It is also used if a cell has exhausted its given text content, and has been
split over a page break. This can happen if other cells in the row have much
more text content than this cell. Therefore, it might be a good idea to I<not>
use a default such as "no cell content", as this could be confusing to readers
who have seen content for this cell on the previous page.

Note that C<max_word_length> splitting is I<not> applied to the default text,
so be careful about using long words.

If you want different effects for different rows, columns, or cells, you can
override the global default setting. If you want B<(No content)> for the first
printout of a cell (split over two or more pages) and B<-> for second and later
printouts, you could leave the global default as '-' and simply give the cell
the text "(No content)" (no longer really I<empty>, but you get the idea). If
you have some content for a cell, and want B<no further entry> after it runs
out of content, you could set C<default_text> to B<no further entry>, and so on.

B<Value:> any string (can be a blank ' ' or an empty string '').

B<Default:> '-'  ($empty_cell_text)

=item B<max_word_length> - Breaks long words 

It may be necessary to break up long words (like serial numbers, hashes, 
etc.) to fit within a column, by adding a space after every Nth symbol, 
unless a space (x20) is found already in the text. 

B<Note> that this does I<not> add a hyphen (dash)!
It merely ensures that there will be no runs of non-space characters longer
than I<N> characters, reducing the chance of forcing an overly wide column.

B<Value:> can be any positive integer number (character count)

B<Default:> C<20>

    'max_word_length' => 25,    # Will add a space after every 25 symbols
                                # unless there is a natural break (space)

=item B<padding> - Padding applied to every cell

=item B<padding_top>    - top cell padding, overrides 'padding'

=item B<padding_right>  - right cell padding, overrides 'padding'

=item B<padding_left>   - left cell padding, overrides 'padding'

=item B<padding_bottom> - bottom padding, overrides 'padding'

B<Value:> can be any non-negative number (E<ge> 0)

B<Default padding:> C<2>.  ($padding_default)

See L</COMPATIBILITY> for returning to the old value of C<0>.

B<Default padding_*> C<'padding'>

    'padding'        => 5,     # all sides cell padding
    'padding_top'    => 8,     # top cell padding, overrides 'padding'
    'padding_right'  => 6,     # right cell padding, overrides 'padding'
    'padding_left'   => 2,     # left cell padding, overrides 'padding'
    'padding_bottom' => undef, # bottom padding will be 5, as it will fall
                               # back to 'padding' value

=item B<font> - instance of PDF::Builder::Resource::Font defining the font to 
be used in the table (or a subsection of it).

B<Value:> can be any PDF::Builder::Resource::* type of font

B<Default:> C<'Times'> core font with I<latin1> encoding

    'font' => $pdf->corefont("Helvetica", -encoding => "latin1"),

B<CAUTION:> Only TrueType and OpenType fonts (ttfont call) can make use of
multibyte encodings such as 'utf8'. Errors will result if you attempt to use
'utf8', etc. with corefont, psfont, etc. font types! For these, you I<must>
only specify a single-byte encoding.

=item B<font_size> - Size of the font that will be used in the table (or a
subsection of it).

B<Value:> can be any positive number

B<Default:> C<12>  ($font_size_default)

    'font_size' => 16,

=item B<fg_color> - Font color for all text.

=item B<bg_color> - Background color for all text.

B<Value:> Color specifier as 'name' or '#rrggbb' (or other suitable color 
specification format)

B<Default:> C<'black'> text on (transparent) background. In other words, there
is no default background color. The exception is for any B<header> row, where 
the default colors are C<#000066> (dark blue, $h_fg_color_default) on 
C<#FFFFAA> (light yellow, $h_bg_color_default).

    'fg_color'      => '#333333',

B<Deprecated names:> I<font_color, background_color>
(both will go away in the future!)

=item B<fg_color_odd> - Font color for odd rows (override C<fg_color>).

=item B<fg_color_even> - Font color for even rows (override C<fg_color>).

=item B<bg_color_odd> - Background color for odd rows (override C<bg_color>).

=item B<bg_color_even> - Background color for even rows (override C<bg_color>).

B<Value:> Color specifier as 'name' or '#rrggbb' (or other suitable color 
specification format)

    'fg_color_odd'  => 'purple',
    'fg_color_even' => '#00FF00',
    'bg_color_odd'  => 'gray',
    'bg_color_even' => 'lightblue',

B<Deprecated names:> I<font_color_odd, font_color_even, 
background_color_odd, background_color_even> (all will go away in the future!)

Note that *_color_odd/even usually make the most sense as global settings,
although it I<is> possible to use them within columns (see chess.pl example),
and even rows and cells, but not header rows.

=item B<underline> - Underline specifications for text in the table.

B<Value:> 'auto', integer of distance (below baseline), or arrayref of 
distance & thickness (more than one pair will provide multiple underlines). 
Negative distance gives strike-through. C<[]> ('none' also works for 
PDF::Builder) gives no underline.

Note that it is unwise to underline all content in the table! It should be
used selectively to I<emphasize> important text, such as header content, or
certain cells. Unfortunately, there is currently no way to turn underlining
off and on I<within> a cell.

B<Default:> none

B<Deprecated name:> I<font_underline> (will go away in the future!)

=item B<min_rh> - Desired minimum row height.

This setting will be honored only if 
C<min_rh E<gt> font_size + padding_top + padding_bottom> (i.e., it is
taller than the calculated minimum value).

This setting doesn't usually make sense when used in a column_props or a
cell_props, but it I<is> possible to do, and may be useful in certain
situations.

B<Value:> can be any positive number

B<Default:> C<font_size + padding_top + padding_bottom>

    'min_rh' => 24,

B<Deprecated name:> I<row_height> (will go away in the future!)

=item B<justify> - Alignment of text in a cell.

B<Value:> One of 'left', 'right', 'center'

B<Default:> C<'left'>

=item B<min_w> - Minimum width of this cell or column. 

PDF::Table will set a cell (and the column it's in) minimum width to fit the
longest word (after splitting on C<max_word_length>) found in the text. This
amount may be increased to C<min_w>. A column should be no narrower than its
widest minimum width, but could be larger in order to fill out the table width.

B<Value:> can be any number satisfying C<0 < min_w < w>

B<Default:> Auto calculated

Note that C<min_w> is usually used for a column_props to set the column
minimum width. If used in a row_props, it will act as a I<global> setting; if
used in a cell_props, that will force the minimum width for the cell's column.

=item B<max_w> - Maximum width of this column. 

PDF::Table will set a cell (and the column it's in) maximum width to fit the
total length of the text content. This will seldom be actually used, but 
C<max_w> may be used to I<reduce> this maximum. When columns are being widened
in order to meet the desired table width, it will try to honor the maximum
width setting and avoid adding any width to a column already at its maximum
width (but this cannot be guaranteed).

B<Value:> can be any number satisfying C<0 < min_w E<le> max_w < w>

B<Default:> Auto calculated

=item B<rule_w> - Width of table rule lines (internal table dividers).

=item B<h_rule_w> - Width of horizontal rules (bottom of a cell).
Overrides 'rule_w' value for horizontal usage.

=item B<v_rule_w> -  Width of vertical rules (left side of a cell). 
Overrides 'rule_w' value for vertical usage.

B<Value:> can be any positive number. When set to 0, it will disable 
rules. This is the line thickness for drawing a rule.

B<Default:> C<1>  (corresponding border value)

A I<rule> is a line bordering a I<cell> in the table. While it does not enter 
into table height or width calculations, be sure to set your C<padding> 
settings to allow sufficient clearance of cell content, especially if you make 
the rules more than a Point or two in thickness (line width). Note that a
cell only defines and draws its left and bottom rules -- the top rule is 
defined in the cell or row above, and the right rule is defined in the cell
or column to the right of this one.

    'rule_w'     => 3,     # rule width is 3
    'h_rule_w'   => 1,     # horizontal rules will be 1, overriding 3
    'v_rule_w'   => undef, # vertical rules will be 3, as it will 
                           # fall back to 'rule_w'

Note that both borders and rules overlay the exact boundary between two cells
(i.e., the centerline). That is, one half of a rule or border will overlay the
adjoining cells. Rules do not expand the size of the table. If you set 
particularly thick (wide) rules, pay attention to adding some padding on the 
appropriate side(s), so that valuable content is not overlaid. For cells along 
the outer border, a I<border> will be drawn instead of a I<rule>.

Cell rules inherit thickness and color from the border settings, so if you want
no internal rules, you need to set 

    'rule_w'     => 0,     # no rules
    
=item B<rule_c> -  Rule color for all rules.

=item B<h_rule_c> -  Rule color for horizontal (bottom) rules, overriding C<rule_c> for this usage.

=item B<v_rule_c> -  Rule color for vertical (left) rules, overriding C<rule_c> for this usage.

B<Value:> Color specifier as 'name' or '#rrggbb'

B<Default:> C<'black'> (corresponding border value)

    'rule_c' => 'red',

=back

=head4 New Page Function Hook

B<new_page_func> is a CODE reference to a function that returns a 
PDF::Builder::Page instance.

If used, the parameter 'C<new_page_func>' must be a function reference which, 
when executed, will create a new page and will return the object to the module.
For example, you can use it to put Page Title, Page Frame, Page Numbers and 
other content that you need.
Also if you need a different paper size and orientation than the default 
US-Letter, e.g., B2-Landscape, you can use this function ref to set it up for 
you. For more info about creating pages, refer to PDF::Builder PAGE METHODS 
Section.
Don't forget that your function must return a page object created with the
PDF::Builder page() method. C<$code_ref> can be something like C<\&new_page>.

    'new_page_func'  => $code_ref,

The C<$code_ref> may be an inline sub definition (as show below), or a regular
named C<sub> (e.g., 'new_page()') referenced as C<\&new_page>. The latter may 
be cleaner than inlining, if the routine is quite long.

An example of reusing a saved PDF page as a I<template>:

    my $pdf      = PDF::API2->new();
    my $template = PDF::API2->open('pdf/template.pdf');
    my $new_page_func = sub { return $pdf->import_page($template, 1); }

    table(
        ...
        new_page_func => $new_page_func,
        ...

This will call a function to grab a copy of a template PDF's page 1 and
insert it as the new last page of the PDF, as the starting point for the next
I<overflow> (continuation) page of the table, if needed. Note that the
C<$template-E<gt>openpage(1)> call is B<unsuitable> for this purpose, as it does
not insert the page into the current PDF.

You can also create a blank page and prefill it with desired content:

    my $pdf      = PDF::API2->new();
    my $new_page_func = sub { 
        my $page = $pdf->page(); # so far, no difference from default behavior
        $page->mediaBox(...);  # set page size/orientation, etc.
        my $text = $page->text();
        # set font, placement, etc.
        $text->text(...);  # write header, footer, etc.
        ...
        return $page;
    }

    table(
        ...
        new_page_func => $new_page_func,
        ...

If C<new_page_func> is not defined, PDF::Table will simply call 
C<$pdf-E<gt>page()> to generate a new, blank, "next" page.

Note that this function is B<not> called for the first page of a table. That 
one uses the current C<$page> parameter passed to the C<table()> call. It is
only called when needed for overflow (C<next_y> and C<next_h>) pages, where
it replaces the C<$page> parameter with a new page framework. You may 
want to consider using the same function to create your other (non-table) 
pages, assuming you want the same format (PDF content) across all pages of the
table.

=head4 Cell Render Hook

B<cell_render_hook> is a CODE reference to a function called with the 
current cell coordinates. If used, the parameter C<cell_render_hook> must be a 
function reference. It is most useful for creating special items within a
text block, such as a URL link inside of a cell. 
The following example adds a link in the first column of each non-header row:

    'cell_render_hook'  => sub {
        my ($page, $first_row, $row, $col, $x, $y, $w, $h) = @_;

        # Do nothing except for first column (and not a header row)
        return unless ($col == 0);
        return if ($first_row);

        # Create link
        my $value = $list_of_vals[$row-1];
        my $url = "https://${hostname}/app/${value}";

        my $annot = $page->annotation();
        $annot->url( $url, -rect => [$x, $y, $x+$w, $y+$h] );
    },

=head4 Header Row Properties

If the 'header_props' parameter is used, it should be a hashref. Passing an 
empty HASH will trigger a header row initialized with Default values.
There is no 'data' variable for the content, because the module assumes that the
first table row will become the header row. It will copy this row and put it on 
every new page if the 'repeat' parameter is set.

=over

=item B<repeat> - Flag showing if header row should be repeated on every new 
page.

B<Value:> 0,1   1-Yes/True, 0-No/False

B<Default:> C<1> ($repeat_default)

See L</COMPATIBILITY> if you wish to change it back to the old behavior 
of C<0>.

    my $hdr_props = {
        'font'       => $pdf->corefont("Helvetica", -encoding => "latin1"),
        'font_size'  => 18,
        'fg_color'   => '#004444',
        'bg_color'   => 'yellow',
        'repeat'     => 0,
        'justify'    => 'center',
    };

=back

=head4 Row Properties

If the 'row_props' parameter is used, it should be an arrayref of hashrefs,
with one hashref for each row of the table. The rows are counted from 
top to bottom, so the hash reference at C<$row_props[0]> will hold properties 
for the first row (from top to bottom).
If you DO NOT want to give properties for a row, but to give for another, 
just insert an empty hash reference into the array for the row that you want 
to skip. This will cause the counting to proceed as expected and the properties 
to be applied at the right rows.

Each hashref can contain any of the keys shown below:

=over

Example:

    my $row_props = [
        # This is an empty hash to indicate default properties for first row
        {},
        # the next hash will hold the properties for the second row from 
        # top to bottom.
        {
            'min_rh'    => 75,        # Minimum row height of 75
            'justify'   => 'right',   # Right text alignment
            'font'      => $pdf->corefont("Helvetica", 
                                          -encoding => "latin1"),
            'font_size' => 10,
            'fg_color'  => 'blue',
            'bg_color'  => '#FFFF00',
        },
        # etc.
    ];

There are no settings unique to rows. Do be aware of when "row 0" may refer
to I<header> row properties!

=back

=head4 Column Properties

If the 'column_props' parameter is used, it should be an arrayref of hashrefs,
with one hashref for each column of the table. The columns are counted from 
left to right, so the hash reference at C<$col_props[0]> will hold properties 
for the first column (from left to right).
If you DO NOT want to give properties for a column, but to give for another, 
just insert an empty hash reference into the array for the column that you want 
to skip. This will cause the counting to proceed as expected and the properties 
to be applied at the right columns.

Each hashref can contain any of the keys shown below:

=over

Example:

    my $col_props = [
        # This is an empty hash to indicate default properties for first col.
        {},
        # the next hash will hold the properties for the second column from 
        # left to right.
        {
            'min_w'     => 100,       # Minimum column width of 100
            'max_w'     => 150,       # Maximum column width of 150
            'justify'   => 'right',   # Right text alignment
            'font'      => $pdf->corefont("Helvetica", 
                                          -encoding => "latin1"),
            'font_size' => 10,
            'fg_color'  => 'blue',
            'bg_color'  => '#FFFF00',
        },
        # etc.
    ];

There are no settings unique to columns.

=back

NOTE: If 'min_w' and/or 'max_w' parameter is used in 'col_props', keep in mind 
that it may be overridden by the calculated minimum/maximum cell width so that 
the table can be created.
When this happens, a warning will be issued with some suggestions on what can 
be done.
In cases of a conflict between column formatting and odd/even row formatting, 
'col_props' will override odd/even.

=head4 Cell Properties

If the 'cell_props' parameter is used, it should be an arrayref with arrays of 
hashrefs (of the same dimension as the data array) with one hashref for each 
cell of the table.

Each hashref can contain any of the keys shown below:

=over

=item B<colspan> - Span this cell over multiple columns to the right.

B<Value:> can be any positive number less than the number of columns to the 
right of the current column

B<Default:> undef

NOTE: If you want to have regular columns B<after> a colspan, you have to 
provide C<undef> for the columns that should be spanned

NOTE: If you use C<colspan> to span a column, but provide data for it, your 
table will be mangled: the spanned-but-data-provided-column will be rendered! 
But, as HTML works the same way, we do not consider this a bug.

Example:

    # row0 col1 should span 2 cols:
    @data = ( [ 'r1c1', 'r1c2', 'r1c3' ], ['r2c1+',undef,'r2c3'] );
    # note: ignoring return values array
    $tab->table( $pdf, $page, \@data, %TestData::required,
      'cell_props' => [
          [],
          [{'colspan' => 2}]
      ]
    );

=back

See the file C<examples/colspan.pl> for detailed usage.

Example:

    my $cell_props = [
        [ # This array is for the first row (0). 
          # If header_props is defined, it will override these settings.
            {    # Row 0 cell 0
                'bg_color'  => '#AAAA00',
                'fg_color'  => 'yellow',
                'underline' => [ 2, 2 ],
            },

            # etc.
        ],
        [ # Row 1 (first data row, if header_props given)
            {    # Row 1 cell 0
                'bg_color' => '#CCCC00',
                'fg_color' => 'blue',
            },
            {    # Row 1 cell 1
                'bg_color' => '#BBBB00',
                'fg_color' => 'red',
            },
            # etc.
        ],
        [ # Row 2
            {    # Row 2 cell 0 span cell 1
                'colspan' => 2
            },
            # etc.
        ],
        # etc.
    ];

    OR

    my $cell_props = [];
    $cell_props->[1][0] = {
        # Row 2 cell 1
        'bg_color' => '#CCCC00',
        'fg_color' => 'blue',
    };

=head2 text_block()

    my ($width_of_last_line, $ypos_of_last_line, $left_over_text) = 
        text_block( $txt, $data, %settings)

=over

=item Description

Utility method to create a block of text. The block may contain multiple 
paragraphs (input C<$data> separated by implicit or explicit newlines C<\n>).
It is mainly used internally, but you can use it from outside for placing 
formatted text anywhere on the sheet.

NOTE: This method will NOT add more pages to the PDF instance if the space is 
not enough to place the string inside the block.
Leftover text will be returned and has to be handled by the caller - i.e., add 
a new page and a new block with the leftover.

=item Parameters

    $txt  - a PDF::Builder::Page::Text instance representing the text tool.
    $data - a string that will be placed inside the block, broken up into
            lines that fit within the indicated width.
    %settings - HASH with geometry and formatting parameters. Note that
                several parameters are mandatory.

=item Returns

The return value is a 3 item list where

    $width_of_last_line - Width of last line in the block
    $final_y - The Y coordinate of the block bottom so that additional 
               content can be added after it
    $left_over_text - Text that did not fit in the provided box geometry.

=item Example

    # PDF::Builder objects
    my $page = $pdf->page();
    my $txt  = $page->text();

    my %settings = (
        # MANDATORY position and table size
        'x' => 10,
        'y' => 570,
        'w' => 220,
        'h' => 180,

        # OPTIONAL PARAMETERS
        'leading'  => $font_size*1.15 | $distance_between_lines,
        'align'    => "left|right|center|justify|fulljustify",
                        default: left
        'max_word_length' => $optional_max_word_chars_between_splits
                        default: 20
        'parspace' => $optional_vertical_space_before_paragraph,
                        default: 0 extra vertical space

        # Only one of the following parameters can be given.
        # They override each other, in the order given. C<hang> is the 
        # highest weight.
        'hang'     => $optional_hanging_text_to_lead_a_paragraph,
        'flindent' => $optional_indent_of_first_line,
        'fpindent' => $optional_indent_of_first_paragraph,
        'indent'   => $optional_indent_of_text_to_every_non_first_line,
    );

    my ( $width_of_last_line, $final_y, $left_over_text ) = 
         $pdftable->text_block( $txt, $data, %settings );

=back

=head1 VERSION

1.005

=head1 AUTHOR

Daemmon Hughes

=head1 DEVELOPMENT

Further development Versions 0.02 -- 0.11 - Desislav Kamenov

Further development since Ver: 0.12 - Phil Perry

=head1 COPYRIGHT AND LICENSE

Copyright (C) 2006 by Daemmon Hughes, portions Copyright (C) 2004 Stone
Environmental Inc. (www.stone-env.com) All Rights Reserved.

Copyright (C) 2020-2024 by Phil M Perry.

This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself, either Perl version 5.8.7 or,
at your option, any later version of Perl 5 you may have available.
Note that Perl 5.10 is the minimum supported level.

=head1 PLUGS

=over

=item by Daemmon Hughes

Much of the original development work on this module was sponsored by
Stone Environmental Inc. (www.stone-env.com).

The text_block() method is a slightly modified copy of the one from
Rick Measham's PDF::API2 tutorial at
http://pdfapi2.sourceforge.net/cgi-bin/view/Main/YourFirstDocument

=item by Desislav Kamenov (@deskata on Twitter)

The development of this module was supported by SEEBURGER AG (www.seeburger.com) till year 2007

Thanks to my friends Krasimir Berov and Alex Kantchev for helpful tips and QA during development of versions 0.9.0 to 0.9.5

Thanks to all GitHub contributors!

=back

=head1 CONTRIBUTION

PDF::Table is on GitHub. You are more than welcome to contribute!

https://github.com/PhilterPaper/PDF-Table

=head1 SEE ALSO

L<PDF::API2>, L<PDF::Builder>

=cut