package String::Truncate;

use warnings;
use strict;

use Carp qw(croak);
use Sub::Install qw(install_sub);

=head1 NAME

String::Truncate - a module for when strings are too long to be displayed in...

=head1 VERSION

version 0.102

 $Id: /my/cs/projects/String-Truncate/trunk/lib/String/Truncate.pm 32171 2007-07-23T21:45:08.267438Z rjbs  $

=cut

our $VERSION = '0.102';

=head1 SYNOPSIS

This module handles the simple but common problem of long strings and finite
terminal width.  It can convert:

 "this is your brain" -> "this is your ..."
                      or "...is your brain"
                      or "this is... brain"
                      or "... is your b..."

It's simple:

 use String::Truncate qw(elide);

 my $brain = "this is your brain";

 elide($brain, 16); # first option
 elide($brain, 16, { truncate => 'left' });   # second option
 elide($brain, 16, { truncate => 'middle' }); # third option
 elide($brain, 16, { truncate => 'ends' });   # fourth option

 String::Trunc::trunc($brain, 16); # => "this is your bra"

=head1 THE BASICS

=head2 C< elide($string, $length, \%arg) >

This function returns the string, if it is less than or equal to C<$length>
characters long.  If it is longer, it truncates the string and marks the
elision.

Valid arguments are:

 elide  - elide at left, right, middle, or ends? (default: right)
 marker - how to mark the elision (default: ...)
 at_space - if true, strings will be broken at whitespace if possible

=cut

my %elider_for = (
  right  => \&_elide_right,
  left   => \&_elide_left,
  middle => \&_elide_middle,
  ends   => \&_elide_ends,
);

sub _elide_right {
  &_assert_1ML; ## no critic Ampersand
  my ($string, $length, $marker, $at_space) = @_;
  my $keep = $length - length($marker);

  if ($at_space) {
    
    my ($substr) = $string =~ /\A(.{0,$keep})\s/;
    $substr = substr($string, 0, $keep) 
      unless defined $substr and length $substr;

    return $substr . $marker;
  } else {
    return substr($string, 0, $keep) . $marker;
  }
}

sub _elide_left {
  &_assert_1ML; ## no critic Ampersand
  my ($string, $length, $marker, $at_space) = @_;
  my $keep = $length - length($marker);
  return $marker
       . reverse(_elide_right(scalar reverse($string), $keep, q{}, $at_space));
}

sub _elide_middle {
  &_assert_1ML; ## no critic Ampersand
  my ($string, $length, $marker, $at_space) = @_;
  my $keep = $length - length($marker);
  my ($keep_left, $keep_right) = (int($keep / 2)) x 2;
  $keep_left +=1 if ($keep_left + $keep_right) < $keep;
  return _elide_right($string, $keep_left, q{}, $at_space)
       . $marker
       . _elide_left($string, $keep_right, q{}, $at_space)
}

sub _elide_ends {
  &_assert_2ML; ## no critic Ampersand
  my ($string, $length, $marker, $at_space) = @_;
  my $midpoint = int(length($string) / 2);
  my $each = int($length / 2);

  return _elide_left(substr($string, 0, $midpoint), $each, $marker, $at_space)
       . _elide_right(substr($string, -$midpoint), $each, $marker, $at_space)
}

sub _assert_1ML {
  my ($string, $length, $marker) = @_;
  croak "elision marker <$marker> is longer than allowed length $length!"
    if length($marker) > $length;
}

sub _assert_2ML {
  my ($string, $length, $marker) = @_;
  # this should only complain if needed: elide('foobar', 3, {marker=>'...'})
  # should be ok -- rjbs, 2006-02-24
  croak "two elision markers <$marker> are longer than allowed length $length!"
    if (length($marker) * 2) > $length;
}

sub elide {
  my ($string, $length, $arg) = @_;
  $arg = {} unless $arg;
  my $truncate = $arg->{truncate} || 'right';

  croak "invalid value for truncate argument: $truncate"
    unless my $elider = $elider_for{ $truncate };

  # hey, this might be really easy:
  return $string if length($string) <= $length;

  my $marker = defined $arg->{marker} ? $arg->{marker} : '...';
  my $at_space = defined $arg->{at_space} ? $arg->{at_space} : 0;
  
  return $elider->($string, $length, $marker, $at_space);
}
  
=head2 C<< trunc($string, $length, \%arg) >>

This acts just like C<elide>, but assumes an empty marker, so it actually
truncates the string normally.

=cut

sub trunc {
  my ($string, $length, $arg) = @_;
  $arg = {} unless $arg;

  croak "marker may not be passed to trunc()" if exists $arg->{marker};
  $arg->{marker} = q{};

  return elide($string, $length, $arg);
}

=head1 IMPORTING

String::Truncate exports both C<elide> and C<trunc>, and also supports the
Exporter-style ":all" tag.

  use String::Truncate ();        # export nothing
  use String::Truncate qw(elide); # export just elide()
  use String::Truncate qw(:all);  # export both elide() and trunc()
  use String::Truncate qw(-all);  # export both elide() and trunc()

When exporting, you may also supply default values:

  use String::Truncate -all => defaults => { length => 10, marker => '--' };

  # or

  use String::Truncate -all => { length => 10, marker => '--' };

These values affect only the imported version of the functions.  You may pass
arguments as usual to override them, and you may call the subroutine by its
fully-qualified name to get the standard behavior.

=cut

use Sub::Exporter::Util ();
use Sub::Exporter -setup => {
  exports => {
    Sub::Exporter::Util::merge_col(defaults => {
      trunc => sub { trunc_with_defaults($_[2]) },
      elide => sub { elide_with_defaults($_[2]) },
    })
  },
  collectors => [ qw(defaults) ]
};

=head1 BUILDING CODEREFS

The imported builds and installs lexical closures (code references) that merge
in given values to the defaults.  You can build your own closures without
importing them into your namespace.  To do this, use the C<elide_with_defaults>
and C<trunc_with_defaults> routines.

=head2 C< elide_with_defaults >

  my $elider = String::Truncate::elide_with_defaults(\%arg);

This routine, never exported, builds a coderef which behaves like C<elide>, but
uses default values when needed.  All the valud arguments to C<elide> are valid
here, as well as C<length>.

=cut

sub _code_with_defaults {
  my ($code, $skip_defaults) = @_;
  
  sub {
    my $defaults = shift || {};
    my %defaults = %$defaults;
    delete $defaults{$_} for @$skip_defaults;

    my $length = delete $defaults{length};

    sub {
      my $string = $_[0];
      my $length = defined $_[1] ? $_[1] : $length;
      my $arg = { %defaults, (defined $_[2] ? %{ $_[2] } : ()) };

      return $code->($string, $length, $arg);
    }
  }
}

BEGIN {
  install_sub({
    code => _code_with_defaults(\&elide),
    as   => 'elide_with_defaults',
  });
}

=head2 C< trunc_with_defaults >

This routine behaves exactly like elide_with_defaults, with one obvious
exception: it retuns code that works like C<trunc> rather than C<elide>.  If a
C<marker> argument is passed, it is ignored.

=cut

BEGIN {
  install_sub({
    code => _code_with_defaults(\&trunc, ['marker']),
    as   => 'trunc_with_defaults',
  });
}

=head1 SEE ALSO

L<Text::Truncate> does a very similar thing.  So does L<Text::Elide>.

=head1 AUTHOR

Ricardo SIGNES, C<< <rjbs at cpan.org> >>

=head1 BUGS

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

=head1 ACKNOWLEDGEMENTS

Ian Langworth gave me some good advice about naming things.  (Also some bad
jokes.  Nobody wants String::ETOOLONG, Ian.)  Hans Dieter Pearcey suggested
allowing defaults just in time for a long bus ride, and I was rescued from
boredom by that suggestion

=head1 COPYRIGHT & LICENSE

Copyright 2005-2007 Ricardo SIGNES.

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

=cut

1; # End of String::Truncate