feat(search): add ocis search optimize CLI command

changelog/unreleased/enhancement-search-optimize-command.md

@@ -0,0 +1,8 @@
+Enhancement: Add `ocis search optimize` command
+
+Added a new CLI command `ocis search optimize` that compacts the search
+index by merging segments without re-indexing content. This is useful for
+existing instances where the index has accumulated many small segments
+from incremental writes, without needing to run a full re-index.
+
+https://github.com/owncloud/ocis/pull/12136

protogen/proto/ocis/services/search/v0/search.proto

@@ -9,6 +9,7 @@ import "protoc-gen-openapiv2/options/annotations.proto";
 import "google/api/field_behavior.proto";
 import "google/api/annotations.proto";
 import "google/protobuf/field_mask.proto";
+import "google/protobuf/empty.proto";
 
 option (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_swagger) = {
   info: {
@@ -47,6 +48,12 @@ service SearchProvider {
         body: "*"
     };
   }
+  rpc OptimizeIndex(google.protobuf.Empty) returns (google.protobuf.Empty) {
+    option (google.api.http) = {
+        post: "/api/v0/search/optimize-index",
+        body: "*"
+    };
+  }
 }
 
 service IndexProvider {

services/search/pkg/command/optimize.go

@@ -0,0 +1,54 @@
+package command
+
+import (
+	"context"
+	"fmt"
+	"time"
+
+	"github.com/urfave/cli/v2"
+	"go-micro.dev/v4/client"
+	"google.golang.org/protobuf/types/known/emptypb"
+
+	"github.com/owncloud/ocis/v2/ocis-pkg/config/configlog"
+	"github.com/owncloud/ocis/v2/ocis-pkg/service/grpc"
+	"github.com/owncloud/ocis/v2/ocis-pkg/tracing"
+	searchsvc "github.com/owncloud/ocis/v2/protogen/gen/ocis/services/search/v0"
+	"github.com/owncloud/ocis/v2/services/search/pkg/config"
+	"github.com/owncloud/ocis/v2/services/search/pkg/config/parser"
+)
+
+// Optimize is the entrypoint for the optimize command.
+func Optimize(cfg *config.Config) *cli.Command {
+	return &cli.Command{
+		Name:     "optimize",
+		Usage:    "compact the search index by merging segments, without re-indexing content",
+		Category: "index management",
+		Before: func(_ *cli.Context) error {
+			return configlog.ReturnFatal(parser.ParseConfig(cfg))
+		},
+		Action: func(_ *cli.Context) error {
+			traceProvider, err := tracing.GetServiceTraceProvider(cfg.Tracing, cfg.Service.Name)
+			if err != nil {
+				return err
+			}
+			grpcClient, err := grpc.NewClient(
+				append(grpc.GetClientOptions(cfg.GRPCClientTLS),
+					grpc.WithTraceProvider(traceProvider),
+				)...,
+			)
+			if err != nil {
+				return err
+			}
+
+			c := searchsvc.NewSearchProviderService("com.owncloud.api.search", grpcClient)
+			_, err = c.OptimizeIndex(context.Background(), &emptypb.Empty{},
+				func(opts *client.CallOptions) { opts.RequestTimeout = 10 * time.Minute })
+			if err != nil {
+				fmt.Println("failed to optimize index: " + err.Error())
+				return err
+			}
+			fmt.Println("index optimization complete")
+			return nil
+		},
+	}
+}

services/search/pkg/command/root.go

@@ -17,6 +17,7 @@ func GetCommands(cfg *config.Config) cli.Commands {
 
 		// interaction with this service
 		Index(cfg),
+		Optimize(cfg),
 
 		// infos about this service
 		Health(cfg),

services/search/pkg/search/service.go

@@ -58,6 +58,7 @@ var (
 type Searcher interface {
 	Search(ctx context.Context, req *searchsvc.SearchRequest) (*searchsvc.SearchResponse, error)
 	IndexSpace(rID *provider.StorageSpaceId) error
+	OptimizeIndex(ctx context.Context) error
 	TrashItem(rID *provider.ResourceId)
 	UpsertItem(ref *provider.Reference)
 	UpdateTags(ref *provider.Reference)
@@ -501,6 +502,22 @@ func (s *Service) IndexSpace(spaceID *provider.StorageSpaceId) error {
 	return nil
 }
 
+// OptimizeIndex compacts the search index segments without re-indexing content.
+func (s *Service) OptimizeIndex(ctx context.Context) error {
+	opt, ok := s.engine.(engine.Optimizer)
+	if !ok {
+		return fmt.Errorf("search engine does not support optimization")
+	}
+
+	logDocCount(s.engine, s.logger)
+	s.logger.Info().Msg("optimizing search index")
+	if err := opt.Optimize(ctx); err != nil {
+		return fmt.Errorf("index optimization failed: %w", err)
+	}
+	s.logger.Info().Msg("index optimization complete")
+	return nil
+}
+
 // TrashItem marks the item as deleted.
 func (s *Service) TrashItem(rID *provider.ResourceId) {
 	err := s.engine.Delete(storagespace.FormatResourceID(rID))