From 9e9f172f8196d5605f048541c1f34c41516563a0 Mon Sep 17 00:00:00 2001
From: Marty Schoch <marty.schoch@gmail.com>
Date: Sun, 11 Sep 2016 12:55:03 -0400
Subject: [PATCH] speed up english possessive filter

previous impl always did full utf8 decode of rune
if we assume most tokens are not possessive this is unnecessary
and even if they are, we only need to chop off last to runes
so, now we only decode last rune of token, and if it looks like
s/S then we proceed to decode second to last rune, and then
only if it looks like any form of apostrophe, do we make any
changes to token, again by just reslicing original to chop
off the possessive extension
---
 analysis/language/en/possessive_filter_en.go  | 18 +++----
 .../language/en/possessive_filter_en_test.go  | 51 +++++++++++++++++++
 2 files changed, 59 insertions(+), 10 deletions(-)

diff --git a/analysis/language/en/possessive_filter_en.go b/analysis/language/en/possessive_filter_en.go
index f322c04a..9bae66f8 100644
--- a/analysis/language/en/possessive_filter_en.go
+++ b/analysis/language/en/possessive_filter_en.go
@@ -10,7 +10,7 @@
 package en
 
 import (
-	"bytes"
+	"unicode/utf8"
 
 	"github.com/blevesearch/bleve/analysis"
 	"github.com/blevesearch/bleve/registry"
@@ -40,15 +40,13 @@ func NewPossessiveFilter() *PossessiveFilter {
 
 func (s *PossessiveFilter) Filter(input analysis.TokenStream) analysis.TokenStream {
 	for _, token := range input {
-		runes := bytes.Runes(token.Term)
-		if len(runes) >= 2 {
-			secondToLastRune := runes[len(runes)-2]
-			lastRune := runes[len(runes)-1]
-			if (secondToLastRune == rightSingleQuotationMark ||
-				secondToLastRune == apostrophe ||
-				secondToLastRune == fullWidthApostrophe) &&
-				(lastRune == 's' || lastRune == 'S') {
-				token.Term = analysis.TruncateRunes(token.Term, 2)
+		lastRune, lastRuneSize := utf8.DecodeLastRune(token.Term)
+		if lastRune == 's' || lastRune == 'S' {
+			nextLastRune, nextLastRuneSize := utf8.DecodeLastRune(token.Term[:len(token.Term)-lastRuneSize])
+			if nextLastRune == rightSingleQuotationMark ||
+				nextLastRune == apostrophe ||
+				nextLastRune == fullWidthApostrophe {
+				token.Term = token.Term[:len(token.Term)-lastRuneSize-nextLastRuneSize]
 			}
 		}
 	}
diff --git a/analysis/language/en/possessive_filter_en_test.go b/analysis/language/en/possessive_filter_en_test.go
index e434bf5d..334d6326 100644
--- a/analysis/language/en/possessive_filter_en_test.go
+++ b/analysis/language/en/possessive_filter_en_test.go
@@ -45,6 +45,12 @@ func TestEnglishPossessiveFilter(t *testing.T) {
 				&analysis.Token{
 					Term: []byte("m"),
 				},
+				&analysis.Token{
+					Term: []byte("s"),
+				},
+				&analysis.Token{
+					Term: []byte("'s"),
+				},
 			},
 			output: analysis.TokenStream{
 				&analysis.Token{
@@ -68,6 +74,12 @@ func TestEnglishPossessiveFilter(t *testing.T) {
 				&analysis.Token{
 					Term: []byte("m"),
 				},
+				&analysis.Token{
+					Term: []byte("s"),
+				},
+				&analysis.Token{
+					Term: []byte(""),
+				},
 			},
 		},
 	}
@@ -84,3 +96,42 @@ func TestEnglishPossessiveFilter(t *testing.T) {
 		}
 	}
 }
+
+func BenchmarkEnglishPossessiveFilter(b *testing.B) {
+
+	input := analysis.TokenStream{
+		&analysis.Token{
+			Term: []byte("marty's"),
+		},
+		&analysis.Token{
+			Term: []byte("MARTY'S"),
+		},
+		&analysis.Token{
+			Term: []byte("marty’s"),
+		},
+		&analysis.Token{
+			Term: []byte("MARTY’S"),
+		},
+		&analysis.Token{
+			Term: []byte("marty＇s"),
+		},
+		&analysis.Token{
+			Term: []byte("MARTY＇S"),
+		},
+		&analysis.Token{
+			Term: []byte("m"),
+		},
+	}
+
+	cache := registry.NewCache()
+	stemmerFilter, err := cache.TokenFilterNamed(PossessiveName)
+	if err != nil {
+		b.Fatal(err)
+	}
+	b.ResetTimer()
+
+	for i := 0; i < b.N; i++ {
+		stemmerFilter.Filter(input)
+	}
+
+}