X-Git-Url: http://git.shiar.net/perl/list-index.git/blobdiff_plain/d6c30ef4ef400982bc1b08e9b7bf93993ea29be5..1c331324738bb8ca05889f1a7baa5e1b40595d38:/lib/List/Index.pm diff --git a/lib/List/Index.pm b/lib/List/Index.pm index c42db44..4a21189 100644 --- a/lib/List/Index.pm +++ b/lib/List/Index.pm @@ -23,31 +23,53 @@ sub ranges { my $pages = $options->{pages } || 1 + int $#$self / $pagesize; $pagesize = @$self / $pages; - my $offset = 0; - my @links; + my $offset = $pagesize + .5; + my @links = (''); while ($offset < @$self) { my $link = substr $self->[$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; - $trim++; + if ($offset > $context - 1) { + # take a value slightly before the current offset + my $before = $self->[$offset - $context - 1]; + # see how much of it matches the current link + my $trim = 1; + for my $match (split //, $before) { + scalar $link =~ /\G\Q$match/g or last; + $trim++; + } + # truncate link upto where the earlier value starts to differ + substr($link, $trim) = '' unless $trim > length $link; + } + + if ($offset + $context < $#$self) { + # take a value after the current offset + my $after = $self->[$offset + $context]; + # see how much of it matches the current link + my $trim = 1; + for my $match (split //, $after) { + scalar $link =~ /\G\Q$match/g or last; + $trim++; + } + # use this link if it's shorter + if ($trim < length $link) { + $link = substr $after, 0, $trim; + } } - substr($link, $trim) = '' unless $trim > length $link; } - push @links, [$link]; + push @links, $link; $offset += $pagesize; } + use List::MoreUtils 'uniq'; + @links = uniq @links; for my $i (0 .. $#links - 1) { - my ($link, $lastchar) = $links[$i + 1]->[0] =~ /(.*)(.)/; - $link .= $lastchar eq '.' ? 'a' : chr( ord($lastchar) - 1 ) - unless $lastchar eq 'a'; - $links[$i]->[1] = $link; + my ($link, $lastchar) = $links[$i + 1] =~ /(.*)(.)/; + $link .= $lastchar le 'a' ? '.' : chr( ord($lastchar) - 1 ); + next if $link eq $links[$i] and $i; + $links[$i] .= '-'.$link; } - $links[-1]->[1] = ''; + $links[-1] .= '-'; return \@links; } @@ -56,6 +78,7 @@ sub rangematch { my ($link) = @_; my ($s1, $s2) = $link =~ /([^-]*) - ([^-]*)/x or return qr/^\Q$link/i; + $s1 =~ s/\.$//; my @allow; if (length $s1) { @@ -66,10 +89,14 @@ sub rangematch { my $prefix = ''; my $char; for my $i (0 .. length($s1) - 1) { + my $lasti = $i == length($s1) - 1; $char = substr $s1, $i, 1; my $next = $char; - $next = chr( ord($char) + 1 ) if length $s1 > $i + 1; + # 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; @@ -78,10 +105,23 @@ sub rangematch { 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 .= $char; + $prefix .= $char eq '.' ? '[^a-z]' : $char; } } @@ -97,11 +137,16 @@ sub rangematch { next if $c1 le $char; } } - push @allow, $prefix."(?![$char-$last])" + + if ($char eq '.') { + next if $i < length($s2) - 1; + } + + push @allow, $prefix.'(?!['.($char eq '.' ? 'a' : $char)."-$last])" if $i or $s1 eq ''; } continue { - $prefix .= $char; + $prefix .= $char eq '.' ? '[^a-z]' : $char; } push @allow, $prefix @@ -119,18 +164,18 @@ __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 ', @$_ for @pages; + say "$_" for @pages; use List::Index 'rangematch'; - my $limit = rangematch('b-bmq'); # matches prefix like 'baa'..'bmq' - @results = grep { $limit } @results; + my $limit = rangematch('b-bmq'); # ge 'b' && le 'bmq' + @request = grep { $limit } @values; =head1 DESCRIPTION