set zoekt.latest_commit_date

sourcegraph · Sep 23, 2024 · 0d92ed3 · 0d92ed3
1 parent 5b7cac9
commit 0d92ed3
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 13 deletions.
diff --git a/api.go b/api.go
@@ -635,13 +635,29 @@ func (r *Repository) UnmarshalJSON(data []byte) error {
 		r.ID = uint32(id)
 	}
 
-	// Sourcegraph indexserver doesn't set repo.Rank, so we set it here based on the
-	// latest commit date. Setting it on read instead of during indexing allows us
-	// to avoid a complete reindex.
-	if r.Rank == 0 {
-		// We use the number of months since 1970 as a simple measure of repo freshness
-		// It has the nice property of being stable across re-indexes and restarts.
+	// Sourcegraph indexserver doesn't set repo.Rank, so we set it here. Setting it
+	// on read instead of during indexing allows us to avoid a complete reindex.
+	//
+	// Prefer "latest_commit_date" over "priority" for ranking. We keep priority for
+	// backwards compatibility.
+	if _, ok := repo.RawConfig["latest_commit_date"]; ok {
+		// We use the number of months since 1970 as a simple measure of repo freshness.
+		// It is monotonically increasing and stable across re-indexes and restarts.
 		r.Rank = monthsSince1970(repo.LatestCommitDate)
+	} else if v, ok := repo.RawConfig["priority"]; ok {
+		r.priority, err = strconv.ParseFloat(v, 64)
+		if err != nil {
+			r.priority = 0
+		}
+
+		// Sourcegraph indexserver doesn't set repo.Rank, so we set it here
+		// based on priority. Setting it on read instead of during indexing
+		// allows us to avoid a complete reindex.
+		if r.Rank == 0 && r.priority > 0 {
+			// Normalize the repo score within [0, 1), with the midpoint at 5,000. This means popular
+			// repos (roughly ones with over 5,000 stars) see diminishing returns from more stars.
+			r.Rank = uint16(r.priority / (5000.0 + r.priority) * maxUInt16)
+		}
 	}
 
 	return nil

diff --git a/cmd/zoekt-sourcegraph-indexserver/index.go b/cmd/zoekt-sourcegraph-indexserver/index.go
@@ -116,11 +116,12 @@ func (o *indexArgs) BuildOptions() *build.Options {
 			Name:     o.Name,
 			Branches: o.Branches,
 			RawConfig: map[string]string{
-				"repoid":   strconv.Itoa(int(o.IndexOptions.RepoID)),
-				"priority": strconv.FormatFloat(o.Priority, 'g', -1, 64),
-				"public":   marshalBool(o.Public),
-				"fork":     marshalBool(o.Fork),
-				"archived": marshalBool(o.Archived),
+				"repoid":             strconv.Itoa(int(o.IndexOptions.RepoID)),
+				"priority":           strconv.FormatFloat(o.Priority, 'g', -1, 64),
+				"public":             marshalBool(o.Public),
+				"fork":               marshalBool(o.Fork),
+				"archived":           marshalBool(o.Archived),
+				"latest_commit_date": "1",
 			},
 		},
 		IndexDir:         o.IndexDir,

diff --git a/cmd/zoekt-sourcegraph-indexserver/index_test.go b/cmd/zoekt-sourcegraph-indexserver/index_test.go
@@ -14,12 +14,13 @@ import (
 	"time"
 
 	"github.com/sourcegraph/log/logtest"
-	proto "github.com/sourcegraph/zoekt/cmd/zoekt-sourcegraph-indexserver/protos/sourcegraph/zoekt/configuration/v1"
-	"github.com/sourcegraph/zoekt/ctags"
 	"google.golang.org/grpc"
 	"google.golang.org/protobuf/testing/protocmp"
 	"google.golang.org/protobuf/types/known/timestamppb"
 
+	proto "github.com/sourcegraph/zoekt/cmd/zoekt-sourcegraph-indexserver/protos/sourcegraph/zoekt/configuration/v1"
+	"github.com/sourcegraph/zoekt/ctags"
+
 	"github.com/google/go-cmp/cmp"
 	"github.com/google/go-cmp/cmp/cmpopts"
 
@@ -493,6 +494,7 @@ func TestIndex(t *testing.T) {
 			"git -C $TMPDIR/test%2Frepo.git update-ref HEAD deadbeef",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.archived 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.fork 0",
+			"git -C $TMPDIR/test%2Frepo.git config zoekt.latest_commit_date 1",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.name test/repo",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.priority 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.public 0",
@@ -515,6 +517,7 @@ func TestIndex(t *testing.T) {
 			"git -C $TMPDIR/test%2Frepo.git update-ref HEAD deadbeef",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.archived 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.fork 0",
+			"git -C $TMPDIR/test%2Frepo.git config zoekt.latest_commit_date 1",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.name test/repo",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.priority 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.public 0",
@@ -546,6 +549,7 @@ func TestIndex(t *testing.T) {
 			"git -C $TMPDIR/test%2Frepo.git update-ref refs/heads/dev feebdaed",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.archived 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.fork 0",
+			"git -C $TMPDIR/test%2Frepo.git config zoekt.latest_commit_date 1",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.name test/repo",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.priority 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.public 0",
@@ -593,6 +597,7 @@ func TestIndex(t *testing.T) {
 			"git -C $TMPDIR/test%2Frepo.git update-ref refs/heads/release 12345678",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.archived 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.fork 0",
+			"git -C $TMPDIR/test%2Frepo.git config zoekt.latest_commit_date 1",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.name test/repo",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.priority 0",
 			"git -C $TMPDIR/test%2Frepo.git config zoekt.public 0",