From 761bd1f02cfe73009edf1e31554c991234626ae3 Mon Sep 17 00:00:00 2001
From: Julie Tibshirani <julie.tibshirani@sourcegraph.com>
Date: Tue, 7 May 2024 10:05:37 -0700
Subject: [PATCH] Rename UseKeywordScoring to mention BM25

---
 api.go                | 10 +++++-----
 api_proto.go          |  4 ++--
 build/scoring_test.go |  6 +++---
 eval.go               |  4 ++--
 score.go              |  6 +++---
 shards/shards_test.go |  2 +-
 6 files changed, 16 insertions(+), 16 deletions(-)

diff --git a/api.go b/api.go
index 6d2497d1..192b6a83 100644
--- a/api.go
+++ b/api.go
@@ -946,10 +946,10 @@ type SearchOptions struct {
 	// will be used. This option is temporary and is only exposed for testing/ tuning purposes.
 	DocumentRanksWeight float64
 
-	// EXPERIMENTAL. If true, use keyword-style scoring instead of the default scoring formula.
-	// Currently, this treats each match in a file as a term and computes an approximation to BM25.
-	// When enabled, all other scoring signals are ignored, including document ranks.
-	UseKeywordScoring bool
+	// EXPERIMENTAL. If true, use text-search style scoring instead of the default scoring formula.
+	// The scoring algorithm treats each match in a file as a term and computes an approximation to
+	// BM25. When enabled, all other scoring signals are ignored, including document ranks.
+	UseBM25Scoring bool
 
 	// Trace turns on opentracing for this request if true and if the Jaeger address was provided as
 	// a command-line flag
@@ -1015,7 +1015,7 @@ func (s *SearchOptions) String() string {
 	addBool("Whole", s.Whole)
 	addBool("ChunkMatches", s.ChunkMatches)
 	addBool("UseDocumentRanks", s.UseDocumentRanks)
-	addBool("UseKeywordScoring", s.UseKeywordScoring)
+	addBool("UseBM25Scoring", s.UseBM25Scoring)
 	addBool("Trace", s.Trace)
 	addBool("DebugScore", s.DebugScore)
 
diff --git a/api_proto.go b/api_proto.go
index 7e17b8ca..ccfccb0f 100644
--- a/api_proto.go
+++ b/api_proto.go
@@ -700,7 +700,7 @@ func SearchOptionsFromProto(p *proto.SearchOptions) *SearchOptions {
 		DocumentRanksWeight:    p.GetDocumentRanksWeight(),
 		Trace:                  p.GetTrace(),
 		DebugScore:             p.GetDebugScore(),
-		UseKeywordScoring:      p.GetUseKeywordScoring(),
+		UseBM25Scoring:         p.GetUseKeywordScoring(),
 	}
 }
 
@@ -725,6 +725,6 @@ func (s *SearchOptions) ToProto() *proto.SearchOptions {
 		DocumentRanksWeight:    s.DocumentRanksWeight,
 		Trace:                  s.Trace,
 		DebugScore:             s.DebugScore,
-		UseKeywordScoring:      s.UseKeywordScoring,
+		UseKeywordScoring:      s.UseBM25Scoring,
 	}
 }
diff --git a/build/scoring_test.go b/build/scoring_test.go
index 9bf243fc..7aea8e00 100644
--- a/build/scoring_test.go
+++ b/build/scoring_test.go
@@ -625,9 +625,9 @@ func checkScoring(t *testing.T, c scoreCase, keywordScoring bool, parserType cta
 		defer ss.Close()
 
 		srs, err := ss.Search(context.Background(), c.query, &zoekt.SearchOptions{
-			UseKeywordScoring: keywordScoring,
-			ChunkMatches:      true,
-			DebugScore:        true})
+			UseBM25Scoring: keywordScoring,
+			ChunkMatches:   true,
+			DebugScore:     true})
 		if err != nil {
 			t.Fatal(err)
 		}
diff --git a/eval.go b/eval.go
index cc0e61f2..2b97d054 100644
--- a/eval.go
+++ b/eval.go
@@ -312,12 +312,12 @@ nextFileMatch:
 		finalCands := d.gatherMatches(nextDoc, mt, known, shouldMergeMatches)
 
 		if opts.ChunkMatches {
-			fileMatch.ChunkMatches = cp.fillChunkMatches(finalCands, opts.NumContextLines, fileMatch.Language, opts.DebugScore)
+			fileMatch.ChunkMatches = cp.fillChunkMatches(finalCands, opts.NumContextLines, opts.UseBM25Scoring, fileMatch.Language, opts.DebugScore)
 		} else {
 			fileMatch.LineMatches = cp.fillMatches(finalCands, opts.NumContextLines, fileMatch.Language, opts.DebugScore)
 		}
 
-		if opts.UseKeywordScoring {
+		if opts.UseBM25Scoring {
 			d.scoreFileUsingBM25(&fileMatch, nextDoc, finalCands, opts)
 		} else {
 			// Use the standard, non-experimental scoring method by default
diff --git a/score.go b/score.go
index 115eabd2..25beef5b 100644
--- a/score.go
+++ b/score.go
@@ -39,9 +39,9 @@ func (m *FileMatch) addScore(what string, computed float64, raw float64, debugSc
 	m.Score += computed
 }
 
-func (m *FileMatch) addKeywordScore(score float64, sumTf float64, L float64, debugScore bool) {
+func (m *FileMatch) addBM25Score(score float64, sumTf float64, L float64, debugScore bool) {
 	if debugScore {
-		m.Debug += fmt.Sprintf("keyword-score:%.2f (sum-tf: %.2f, length-ratio: %.2f)", score, sumTf, L)
+		m.Debug += fmt.Sprintf("bm25-score:%.2f (sum-tf: %.2f, length-ratio: %.2f)", score, sumTf, L)
 	}
 	m.Score += score
 }
@@ -160,5 +160,5 @@ func (d *indexData) scoreFileUsingBM25(fileMatch *FileMatch, doc uint32, cands [
 		score += ((k + 1.0) * tf) / (k*(1.0-b+b*L) + tf)
 	}
 
-	fileMatch.addKeywordScore(score, sumTf, L, opts.DebugScore)
+	fileMatch.addBM25Score(score, sumTf, L, opts.DebugScore)
 }
diff --git a/shards/shards_test.go b/shards/shards_test.go
index 5c4ddc73..9b7d37b4 100644
--- a/shards/shards_test.go
+++ b/shards/shards_test.go
@@ -1103,7 +1103,7 @@ func TestUseKeywordScoring(t *testing.T) {
 		&query.Substring{Pattern: "three"})
 
 	opts := zoekt.SearchOptions{
-		UseKeywordScoring: true,
+		UseBM25Scoring: true,
 	}
 
 	results, err := ss.Search(context.Background(), q, &opts)