From 45e9eaaacbdc885efee9c9efefc41ebc92099a92 Mon Sep 17 00:00:00 2001
From: Steve Yen <steve.yen@gmail.com>
Date: Wed, 30 Dec 2015 21:15:24 -0800
Subject: [PATCH] firestorm.indexField() allocs up-front array of TermFreqRow's

This uses the "backing array" technique to allocate many TermFreqRow's
at the front of firestorm.indexField(), instead of the previous
one-by-one, as-needed TermFreqRow allocation approach.

Results from micro-benchmark, null-firestorm, bleve-blast has this
change producing a ~half MB/sec improvement.
---
 index/firestorm/analysis.go |  8 ++++++--
 index/firestorm/termfreq.go | 22 +++++++++++-----------
 2 files changed, 17 insertions(+), 13 deletions(-)

diff --git a/index/firestorm/analysis.go b/index/firestorm/analysis.go
index 4e5b71dc..32c6cf66 100644
--- a/index/firestorm/analysis.go
+++ b/index/firestorm/analysis.go
@@ -97,19 +97,23 @@ func (f *Firestorm) Analyze(d *document.Document) *index.AnalysisResult {
 func (f *Firestorm) indexField(docID []byte, docNum uint64, includeTermVectors bool, fieldIndex uint16, fieldLength int, tokenFreqs analysis.TokenFrequencies) []index.IndexRow {
 
 	rows := make([]index.IndexRow, 0, len(tokenFreqs))
+	tfrs := make([]TermFreqRow, len(tokenFreqs))
+
 	fieldNorm := float32(1.0 / math.Sqrt(float64(fieldLength)))
 
+	i := 0
 	for _, tf := range tokenFreqs {
 		var termFreqRow *TermFreqRow
 		if includeTermVectors {
 			tv, newFieldRows := f.termVectorsFromTokenFreq(fieldIndex, tf)
 			rows = append(rows, newFieldRows...)
-			termFreqRow = NewTermFreqRow(fieldIndex, tf.Term, docID, docNum, uint64(tf.Frequency()), fieldNorm, tv)
+			termFreqRow = InitTermFreqRow(&tfrs[i], fieldIndex, tf.Term, docID, docNum, uint64(tf.Frequency()), fieldNorm, tv)
 		} else {
-			termFreqRow = NewTermFreqRow(fieldIndex, tf.Term, docID, docNum, uint64(tf.Frequency()), fieldNorm, nil)
+			termFreqRow = InitTermFreqRow(&tfrs[i], fieldIndex, tf.Term, docID, docNum, uint64(tf.Frequency()), fieldNorm, nil)
 		}
 
 		rows = append(rows, termFreqRow)
+		i++
 	}
 
 	return rows
diff --git a/index/firestorm/termfreq.go b/index/firestorm/termfreq.go
index 0d5cc9e8..1d36ec26 100644
--- a/index/firestorm/termfreq.go
+++ b/index/firestorm/termfreq.go
@@ -46,18 +46,18 @@ func NewTermVector(field uint16, pos uint64, start uint64, end uint64, arrayPos
 }
 
 func NewTermFreqRow(field uint16, term []byte, docID []byte, docNum uint64, freq uint64, norm float32, termVectors []*TermVector) *TermFreqRow {
-	rv := TermFreqRow{
-		field:  field,
-		term:   term,
-		docID:  docID,
-		docNum: docNum,
-	}
+	return InitTermFreqRow(&TermFreqRow{}, field, term, docID, docNum, freq, norm, termVectors)
+}
 
-	rv.value.Freq = proto.Uint64(freq)
-	rv.value.Norm = proto.Float32(norm)
-	rv.value.Vectors = termVectors
-
-	return &rv
+func InitTermFreqRow(tfr *TermFreqRow, field uint16, term []byte, docID []byte, docNum uint64, freq uint64, norm float32, termVectors []*TermVector) *TermFreqRow {
+	tfr.field = field
+	tfr.term = term
+	tfr.docID = docID
+	tfr.docNum = docNum
+	tfr.value.Freq = proto.Uint64(freq)
+	tfr.value.Norm = proto.Float32(norm)
+	tfr.value.Vectors = termVectors
+	return tfr
 }
 
 func NewTermFreqRowKV(key, value []byte) (*TermFreqRow, error) {