Item.pm

# Copyright (C) 2004, Parrot Foundation.

=head1 NAME

Parrot::Docs::Item - Documentation item

=head1 SYNOPSIS

    use Parrot::Docs::Item;

=head1 DESCRIPTION

A documentation I<item> is one or more related paths with some optional
descriptive text.

Directory paths will be expanded to all the file paths within the
directory and any subdirectories recursively (see the C<files()> method
in C<Parrot::IO::Directory>).

If an item has more than one file associated with it, and has no text,
then an attempt will be made to extract short descriptions from each
file to place under the file path in the index HTML.

=head2 Class Methods

=over

=cut

package Parrot::Docs::Item;

use strict;
use warnings;

use Parrot::Docs::Directory;
use Parrot::Docs::POD2HTML;
use Parrot::Docs::Text2HTML;

=item C<new_item($text, @paths)>

Returns a new item.

Use this when creating items within a C<Parrot::Docs::Section>
subclass's C<new()> method.

=cut

sub new_item {
    my $self = shift;

    return Parrot::Docs::Item->new(@_);
}

=item C<new($text, @contents)>

Returns a new item. If there is no descriptive text then C<$text> should
be an empty string.

The paths in C<@contents> will be interpreted as being relative to the
C<$target> argument in C<write_html()>. There should be at least one
path otherwise an exception is raised.

=cut

sub new {
    my $self     = ref $_[0] ? ref shift : shift;
    my $text     = shift;
    my @contents = @_;

    die "No contents ($text).\n" unless @contents;

    $self = bless {
        TEXT     => $text,
        TITLE    => $text,
        CONTENTS => \@contents,
    }, $self;

    return $self;
}

=back

=head2 Instance Methods

=over 4

=item C<set_parent($parent)>

=item C<parent()>

Accessors for the containing section/group for the item.

=cut

sub set_parent {
    my $self = shift;

    $self->{PARENT} = shift;
}

sub parent {
    my $self = shift;

    return $self->{PARENT};
}

=item C<html_navigation($path)>

Returns the HTML navigation bar.

=cut

sub html_navigation {
    my $self   = shift;
    my $path   = shift;
    my $parent = $self->parent || return '';

    return join ' &raquo; ', grep { length } $parent->html_navigation($path), $parent->html_link($path);
}

=item C<write_html($source, $target, $silent)>

C<$source> is the directory in which the section's contents will be
looked for.

C<$target> is directory into which the section's output will be written.

If C<$silent> is true then progress is not reported.

Any POD-formatted text in the item's files is converted to HTML and
written to a file in C<$target> and an HTML link is created to it.

Alternatively, if a file responds true to C<is_docs_link()> then an HTML
link is created to the file itself.

Some HTML-formatted text describing the files linked to is returned.

=cut

sub write_html {
    my $self       = shift;
    my $source     = shift || die "No source\n";
    my $target     = shift || die "No target\n";
    my $silent     = shift || 0;
    my $index_html = '';
    my @rel_paths  = $self->contents_relative_to_source($source);
    my @short_desc = ();

    foreach my $rel_path (@rel_paths) {
        my $file      = $source->file_with_relative_path($rel_path);

        if ( $file->contains_pod ) {
            print "\n", $rel_path unless $silent;

            my $formatter = Parrot::Docs::POD2HTML->new;
            $formatter->no_errata_section(1); # don't dump errors into HTML output
            $formatter->write_html( $source, $target, $rel_path, $self );

            my $title = $self->{TITLE} || $file->short_description;

            if ($title) {
                $index_html .= $formatter->html_link( $formatter->append_html_suffix($rel_path),
                    $title );
            }
            else {
                $index_html .= $formatter->html_link( $formatter->append_html_suffix($rel_path),
                    $source->relative_path( $file->path ) );
            }

            $index_html .= "<br>\n";

            next if $self->{TEXT};

            my $short_desc = $file->short_description;

            next unless $short_desc;

            next if grep { $_ eq $short_desc } @short_desc;

            push @short_desc, $short_desc;
        }
        elsif ( $file->is_docs_link ) {
            print "\n", $rel_path unless $silent;

            my $formatter = Parrot::Docs::Text2HTML->new;
            $formatter->write_html( $source, $target, $rel_path, $self );

            $index_html .= $formatter->html_link( $formatter->append_html_suffix($rel_path),
                $source->relative_path( $file->path ) );

        }
    }

    return '' unless $index_html;

    if ( $self->{DESCRIPTION} ) {
        $index_html .= "<br>$self->{DESCRIPTION}\n";
    }

    $index_html = '<li>' . $index_html . "</li>\n";

    return $index_html;
}

=item C<contents_relative_to_source($source)>

Returns the contents of the item interpreted relative to the source
directory.

=cut

sub contents_relative_to_source {
    my $self     = shift;
    my $source   = shift;
    my @contents = ();

    foreach my $content ( @{ $self->{CONTENTS} } ) {
        push @contents, $self->file_paths_relative_to_source( $source, $content );

    }

    return @contents;
}

=item C<file_paths_relative_to_source($source, $path)>

If C<$path> is an immediate subdirectory of C<$source>, then this method
returns all the file paths within the directory and any subdirectories
recursively, relative to C<$source>.

If C<$path> is a file in C<$source> then C<$path> is returned.

If C<$path> cannot be found then a warning is printed.

=cut

sub file_paths_relative_to_source {
    my $self      = shift;
    my $source    = shift;
    my $rel_path  = shift;
    my @rel_paths = ();

    if ( $source->relative_path_is_directory($rel_path) ) {
        my $dir = $source->directory_with_relative_path($rel_path);

        # There may be editor scratch files to ignore.

        foreach my $file ( $dir->files( 1, '^\.' ) ) {
            push @rel_paths, $source->relative_path( $file->path );
        }
    }
    elsif ( $source->relative_path_is_file($rel_path) ) {
        push @rel_paths, $rel_path;
    }
    else {
        warn "Failed to process $rel_path.\n";
    }

    return @rel_paths;
}

=back

=head1 SEE ALSO

=over 4

=item C<Parrot::Docs::Section>

=item C<Parrot::Docs::Group>

=back

=cut

1;

# Local Variables:
#   mode: cperl
#   cperl-indent-level: 4
#   fill-column: 100
# End:
# vim: expandtab shiftwidth=4: