X-Git-Url: http://git.shiar.nl/perl/list-index.git/blobdiff_plain/d61e8685992586de85f6b7b58fee20f15c8111b0..e2a186365224032ff37caac824d50c9b43130952:/lib/List/Index.pm
diff --git a/lib/List/Index.pm b/lib/List/Index.pm
index 6a818bb..35f0549 100644
--- a/lib/List/Index.pm
+++ b/lib/List/Index.pm
@@ -6,48 +6,93 @@ use warnings;
use Exporter 'import';
-our $VERSION = '1.00';
+our $VERSION = '1.02';
our @EXPORT_OK = qw(rangematch);
sub new {
- my ($class, $values) = @_;
- bless [sort map { s/[^a-z]/./g; $_ } @$values], $class;
+ my ($class, $options) = @_;
+ $options ||= {};
+ bless $options, $class;
}
sub ranges {
my $self = shift;
+ my @rows = sort map { s/[^a-z]/./g; $_ } @{ shift() };
my $options = shift || {};
+ $options->{$_} //= $self->{$_} for keys %$self;
+
my $pagesize = $options->{pagesize} || 50;
my $context = $options->{context } // 1 + ($pagesize >> 4);
my $length = $options->{length } || 4;
- my $pages = $options->{pages } || 1 + int $#$self / $pagesize;
+ my $pages = $options->{pages } || 1 + int($#rows / $pagesize);
+
+ $pagesize = $pages >= @rows ? 1 : @rows / $pages;
+ my $shrunk = 0;
+ my $enlarged = 0;
- $pagesize = @$self / $pages;
- my $offset = 0;
- my @links;
- while ($offset < @$self) {
- my $link = substr $self->[$offset], 0, $length;
+ my @links = ('');
+ for (my $offset = $pagesize + .5; $offset < @rows; $offset += $pagesize) {
+ my $link = substr $rows[$offset], 0, $length;
if ($context) {
- my $trim = 1;
- my $before = $offset > $context ? $self->[$offset - $context] : '';
- for my $match (split //, $before) {
- scalar $link =~ /\G\Q$match/g or last;
- $trim++;
+ my $lookbehind = -$context + $shrunk;
+ my $lookahead = $context - $enlarged;
+ $shrunk = $enlarged = 0;
+
+ # take a value slightly before the current offset
+ if ((my $before = $offset + $lookbehind) > 0) {
+ # see how much of it matches the current link
+ my $trim = 1;
+ for my $match (split //, $rows[$before - 1]) {
+ scalar $link =~ /\G\Q$match/g or last;
+ $trim++;
+ }
+ # truncate link upto where the earlier value starts to differ
+ if ($trim < length $link) {
+ substr($link, $trim) = '';
+ for (reverse $before .. $offset - 1) {
+ $rows[$_] =~ /^\Q$link/ or last;
+ $enlarged++;
+ }
+ }
+ }
+
+ # take a value after the current offset
+ if ((my $after = $offset + $lookahead) < $#rows) {
+ # see how much of it matches the current link
+ my $trim = 1;
+ pos $link = 0;
+ for my $match (split //, $rows[$after]) {
+ scalar $link =~ /\G\Q$match/g or last;
+ $trim++;
+ }
+ # use this link if it's shorter
+ if ($trim < length $link) {
+ $enlarged = 0;
+ for ($offset + 1 .. $after) {
+ my $prefix = substr $rows[$_], 0, $trim;
+ # advance lookbehind offset on the next page
+ $shrunk++;
+ next if $link =~ /^\Q$prefix/;
+ $link = $prefix;
+ last;
+ }
+ }
}
- substr($link, $trim) = '' unless $trim > length $link;
}
- push @links, [$link];
- $offset += $pagesize;
+ push @links, $link unless $links[-1] eq $link;
}
+ # add range end to each link
for my $i (0 .. $#links - 1) {
- my ($link, $lastchar) = $links[$i + 1]->[0] =~ /(.*)(.)/;
- $link .= $lastchar eq '.' ? 'z' : chr( ord($lastchar) - 1 )
- unless $lastchar eq 'a';
- $links[$i]->[1] = $link;
+ # end at start of next value with the last character decremented
+ my $next = $links[$i + 1];
+ $next =~ s{(.)$}{ $1 le 'a' ? '.' : chr( ord($1) - 1 ) }e;
+ # amend range if it's ahead
+ $links[$i] .= '-'.$next unless $next eq $links[$i];
}
- $links[-1]->[1] = '';
+ # final value takes the rest
+ $links[-1] .= '-';
return \@links;
}
@@ -56,36 +101,80 @@ sub rangematch {
my ($link) = @_;
my ($s1, $s2) = $link =~ /([^-]*) - ([^-]*)/x
or return qr/^\Q$link/i;
+ $s1 =~ s/\.$//;
my @allow;
if (length $s1) {
+ if (length $s2) {
+ $s1 le $s2 or $s1 =~ /^\Q$s2/ or return undef;
+ }
+
my $prefix = '';
- my $c1;
+ my $char;
for my $i (0 .. length($s1) - 1) {
- $c1 = substr $s1, $i, 1;
- my $c2 = length $s2 <= $i ? undef : substr $s2, $i, 1;
- my $next = $i + 1 >= length($s1) ? $c1 : chr( ord($c1) + 1 );
- $next le $c2 or next if defined $c2;
- my $last = defined $c2 && $i == 0 ? chr( ord($c2) - (length $s2 > 1) ) : 'z';
+ my $lasti = $i == length($s1) - 1;
+ $char = substr $s1, $i, 1;
+ my $next = $char;
+ # do not include prefix character in final range
+ $next = chr( ord($char) + 1 ) unless $lasti;
+
+ my $last = 'z';
+ next if $next gt $last;
+ if (length $s2 > $i) {
+ if ($s2 =~ /^\Q$prefix/) {
+ $last = substr $s2, $i, 1;
+ next if $char eq $last;
+ $last = chr( ord($last) - (length $s2 > 1) );
+ next if $next gt $last;
+ }
+ }
+
+ if ($char eq '.') {
+ if ($last eq 'z') {
+# push @allow, $prefix if $i and $lasti;
+# next;
+ }
+# if ($last eq 'z') {
+# push @allow, $prefix if $i and $lasti;
+# next;
+# }
+ $next = 'a';
+ }
+
push @allow, $prefix."[$next-$last]";
}
continue {
- $prefix .= $c1;
+ $prefix .= $char eq '.' ? '[^a-z]' : $char;
}
}
if (length $s2) {
my $prefix = '';
+ my $char;
for my $i (0 .. length($s2) - 1) {
- my $c1 = length $s1 <= $i ? undef : substr $s1, $i, 1;
- my $c2 = substr $s2, $i, 1;
+ $char = substr $s2, $i, 1;
my $last = 'z';
- push @allow, "$prefix(?![$c2-$last])"
+ if (length $s1 > $i) {
+ my $c1 = substr $s1, $i, 1;
+ if ($s1 =~ /^\Q$prefix/) {
+ next if $c1 le $char;
+ }
+ }
+
+ if ($char eq '.') {
+ next if $i < length($s2) - 1;
+ }
+
+ push @allow, $prefix.'(?!['.($char eq '.' ? 'a' : $char)."-$last])"
if $i or $s1 eq '';
- $prefix .= $c2;
}
+ continue {
+ $prefix .= $char eq '.' ? '[^a-z]' : $char;
+ }
+
push @allow, $prefix
- unless length $s1 > length $s2 or length $s1 != 0 && length $s2 == 1; #TODO
+ if $s2 =~ /^\Q$prefix/ and $s1 le $s2
+ and not (length $s2 == 1 && length $s1 >= length $s2 && $s1 ne $s2);
}
my $match = sprintf @allow <= 1 ? '%s' : '(?:%s)', join('|', @allow);
@@ -98,14 +187,18 @@ __END__
=head1 NAME
-List::Index - Paginate alphabetic entries by finding minimal prefixes
+List::Index - Find and apply prefix ranges to paginate keywords
=head1 SYNOPSIS
use List::Index;
- my $index = List::Index->new(\@values);
- my @pages = $index->ranges({pagesize => 50});
- printf '%1$s ', @$_ for @pages;
+ my $index = List::Index->new({ pagesize => 50 });
+ my @pages = $index->ranges(\@values);
+ say "$_" for @pages;
+
+ use List::Index 'rangematch';
+ my $limit = rangematch('b-bmq'); # ge 'b' && le 'bmq'
+ @request = grep { $limit } @values;
=head1 DESCRIPTION