1 files changed, 103 insertions, 0 deletions
diff --git a/lib/search/snippet.js b/lib/search/snippet.js
new file mode 100644
index 0000000..de71911
--- /dev/null
+++ b/lib/search/snippet.js
@@ -0,0 +1,103 @@
+var util = require('../util/util')
+
+function bold_terms (s, terms) {
+  
+}
+sub bold_terms
+	{
+	my ($self, $string, $terms) = @_;
+	$string = $self->strip_html($string);
+	foreach my $term (@$terms)
+		{
+		$string =~ s/\b($term)\b/<b>$1<\/b>/gi;
+		}
+	return $string;
+	}
+sub bold_snippet
+	{
+	my ($self, $string, $terms) = @_;
+	my $snippet = $self->snippet($string, $terms);
+	return $self->bold_terms($snippet, $terms);
+	}
+sub snippet
+	{
+	my ($self, $string, $terms) = @_;
+
+	# clean up the string we got
+	$string = $self->strip_html($string);
+
+	# create a regex out of the search terms
+	my $term_re = join "|", @$terms;
+
+	# take the string to be snippetized and split it into words
+	my @words = split /\s+/, $string;
+
+	# deduper for matching @words indexes, so we don't add a word twice
+	my $index_matches = {};
+
+	# words in the eventual snippet
+	my @words_matched;
+
+	# the snippet itself
+	my $snippet = '';
+
+	# counter for aggregating context after a match
+	my $aggr = 0;
+
+	# amount of context to show, in number of words surrounding a match
+	my $pad = 4;
+
+	# loop over each of the words in the string
+	for (my $i = 0; $i < scalar @words; $i++)
+		{
+		# does this word contain a match?
+		if ($words[$i] =~ /\b($term_re)\b/i && ! $self->is_stopword($1))
+			{
+			# if we aren't already aggregating, add an ellipsis
+			if (! $aggr)
+				{
+				push @words_matched, "...";
+				}
+			# look backward $pad words
+			for (my $j = -$pad; $j < 1; $j++)
+				{
+				# create a new index from the offset
+				my $idx = $i + $j;
+
+				# is this a valid index? has it already been encountered?
+				next if $idx < 0;
+				next if $idx > scalar @words;
+				next if exists $index_matches->{$i+$j};
+
+				# checks out, save this word
+				push @words_matched, $words[$i+$j];
+
+				# note the matching index in our deduper
+				$index_matches->{$i+$j} ++;
+				}
+			# enter aggregate mode -- add the next $pad words
+			$aggr = $pad;
+			}
+		# have we been told to aggregate?
+		elsif ($aggr)
+			{
+			# save this word
+			push @words_matched, $words[$i];
+
+			# add index to the deduper
+			$index_matches->{$i} ++;
+
+			# one less word to aggregate
+			$aggr--;
+			}
+		# keep snippets to a modest length
+		last if scalar @words_matched > 30;
+		}
+	# add a trailing ellipsis
+	push @words_matched, "...";
+
+	# create the snippet from the saved context words
+	$snippet = join " ", @words_matched;
+
+	return $snippet;
+	}
+\ No newline at end of file