Fix/xmp image tags overwrite fix

app/upload/run.go

@@ -6,6 +6,7 @@
 	"fmt"
 	"strings"
 	"sync"
+	"time"
 
 	"github.com/gdamore/tcell/v2"
 	"github.com/simulot/immich-go/adapters"
@@ -54,12 +55,21 @@
 		uc.app.Log().Info("created tag", "tag", tag.Value)
 		tag.ID = r[0].ID
 	}
-	_, err := uc.client.Immich.TagAssets(ctx, tag.ID, ids)
-	if err != nil {
-		uc.app.Log().Error("failed to add assets to tag", "err", err, "tag", tag.Value, "assets", len(ids))
-		return tag, err
+	const batchSize = 500
+	total := len(ids)
+	var err error
+	for i := 0; i < total; i += batchSize {
+		end := i + batchSize
+		if end > total {
+			end = total
+		}
+		_, err = uc.client.Immich.TagAssets(ctx, tag.ID, ids[i:end])
+		if err != nil {
+			uc.app.Log().Error("failed to add assets to tag", "err", err, "tag", tag.Value, "assets", len(ids[i:end]))
+			return tag, err
+		}
 	}
-	uc.app.Log().Info("updated tag", "tag", tag.Value, "assets", len(ids))
+	uc.app.Log().Info("updated tag", "tag", tag.Value, "assets", total)
 	return tag, err
 }
 
@@ -97,16 +107,39 @@
 		return nil
 	}
 	defer func() { uc.finished = true }()
-	// do waiting operations
-	uc.albumsCache.Close()
-	uc.tagsCache.Close()
+
+	if uc.DeferTags {
+		if uc.client.PauseImmichBackgroundJobs {
+			uc.app.Log().Info("Resuming metadata extraction...")
+			// Use a background context to ensure the command is sent even if the main context is cancelling
+			bgCtx := context.Background()
+			_, err := uc.client.AdminImmich.SendJobCommand(bgCtx, "metadataExtraction", "resume", true)
+			if err != nil {
+				uc.app.Log().Error("Failed to resume metadata extraction", "err", err)
+			}
+		}
+
+		uc.app.Log().Info("Waiting for metadata extraction to complete...")
+		err := uc.waitForMetadataExtraction(ctx)
+		if err != nil {
+			uc.app.Log().Error("Failed to wait for metadata extraction", "err", err)
+		}
+		uc.app.Log().Info("Metadata extraction complete, applying tags...")
+		uc.tagsCache.Close()
+	}
 
 	// Resume immich background jobs if requested
 	err := uc.resumeJobs(ctx)
 	if err != nil {
 		return err
 	}
 
+	// do waiting operations
+	uc.albumsCache.Close()
+	if !uc.DeferTags {
+		uc.tagsCache.Close()
+	}
+
 	// Generate FileProcessor report
 	if uc.app.FileProcessor() != nil {
 		report := uc.app.FileProcessor().GenerateReport()
@@ -141,7 +174,11 @@
 	uc.albumsCache = cache.NewCollectionCache(50, func(album assets.Album, ids []string) (assets.Album, error) {
 		return uc.saveAlbum(ctx, album, ids)
 	})
-	uc.tagsCache = cache.NewCollectionCache(50, func(tag assets.Tag, ids []string) (assets.Tag, error) {
+	tagCacheSize := 50
+	if uc.DeferTags {
+		tagCacheSize = 1 << 30
+	}
+	uc.tagsCache = cache.NewCollectionCache(tagCacheSize, func(tag assets.Tag, ids []string) (assets.Tag, error) {
 		return uc.saveTags(ctx, tag, ids)
 	})
 
@@ -596,3 +633,28 @@
 	return nil
 }
 */
+
+func (uc *UpCmd) waitForMetadataExtraction(ctx context.Context) error {
+	ticker := time.NewTicker(2 * time.Second)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		case <-ticker.C:
+			jobs, err := uc.client.Immich.GetJobs(ctx)
+			if err != nil {
+				return err
+			}
+			job, ok := jobs["metadataExtraction"]
+			if !ok {
+				// Job not found, assume it's done or not running
+				return nil
+			}
+			if job.JobCounts.Active == 0 && job.JobCounts.Waiting == 0 {
+				return nil
+			}
+		}
+	}
+}

app/upload/upload.go

@@ -77,6 +77,7 @@ type UpCmd struct {
 	tagsCache         *cache.CollectionCache[assets.Tag]   // List of tags present on the server
 	finished          bool                                 // the finish task has been run
 	infoCollector     *filenames.InfoCollector             // Collects information about the files being processed
+	DeferTags         bool                                 // Defer tagging until metadata extraction is complete
 }
 
 func (uc *UpCmd) RegisterFlags(flags *pflag.FlagSet) {
@@ -85,6 +86,7 @@ func (uc *UpCmd) RegisterFlags(flags *pflag.FlagSet) {
 	flags.BoolVar(&uc.Overwrite, "overwrite", false, "Always overwrite files on the server with local versions")
 	flags.StringSliceVar(&uc.Tags, "tag", nil, "Add tags to the imported assets. Can be specified multiple times. Hierarchy is supported using a / separator (e.g. 'tag1/subtag1')")
 	flags.BoolVar(&uc.SessionTag, "session-tag", false, "Tag uploaded photos with a tag \"{immich-go}/YYYY-MM-DD HH-MM-SS\"")
+	flags.BoolVar(&uc.DeferTags, "defer-tags", false, "Defer tagging until metadata extraction is complete")
 
 	uc.StackOptions.RegisterFlags(flags)
 }

docs/commands/upload.md

@@ -41,11 +41,12 @@ All upload sub-commands require these connection parameters:
 
 ## Tagging and Organization
 
-| Option          | Default      | Description                                  |
-| --------------- | ------------ | -------------------------------------------- |
-| `--session-tag` | `false`      | Tag with upload session timestamp            |
-| `--tag`         | -            | Add custom tags (can be used multiple times) |
-| `--device-uuid` | `$LOCALHOST` | Set device identifier                        |
+| Option          | Default      | Description                                                  |
+| --------------- | ------------ | ------------------------------------------------------------ |
+| `--session-tag` | `false`      | Tag with upload session timestamp                            |
+| `--defer-tags`  | `false`      | Apply tags only after metadata extraction has completed to keep embedded file tags/keywords intact |
+| `--tag`         | -            | Add custom tags (can be used multiple times)                 |
+| `--device-uuid` | `$LOCALHOST` | Set device identifier                                        |
 
 ## User Interface
 
@@ -270,4 +271,4 @@ immich-go upload from-immich \
 
 - [Configuration Options](../configuration.md)
 - [Technical Details](../technical.md)
-- [Best Practices](../best-practices.md)
+- [Best Practices](../best-practices.md)

docs/configuration.md

@@ -150,6 +150,7 @@ overwrite = false
 pause-immich-jobs = true
 server = 'https://immich.app'
 session-tag = false
+defer-tags = false
 skip-verify-ssl = false
 time-zone = ''
 
@@ -473,6 +474,7 @@ upload:
   pause-immich-jobs: true
   server: https://immich.app
   session-tag: false
+  defer-tags: false
   skip-verify-ssl: false
   tag: {}
   time-zone: ""
@@ -710,6 +712,7 @@ upload:
     "pause-immich-jobs": true,
     "server": "https://immich.app",
     "session-tag": false,
+    "defer-tags": false,
     "skip-verify-ssl": false,
     "tag": {},
     "time-zone": ""

docs/environment.md

@@ -165,6 +165,7 @@ The following environment variables can be used to configure `immich-go`.
 | `IMMICH_GO_UPLOAD_PAUSE_IMMICH_JOBS` | `--pause-immich-jobs` | `true` | Pause Immich background jobs during upload operations |
 | `IMMICH_GO_UPLOAD_SERVER` | `--server` |  | Immich server address (example http://your-ip:2283 or https://your-domain) |
 | `IMMICH_GO_UPLOAD_SESSION_TAG` | `--session-tag` | `false` | Tag uploaded photos with a tag "{immich-go}/YYYY-MM-DD HH-MM-SS" |
+| `IMMICH_GO_UPLOAD_DEFER_TAGS` | `--defer-tags` | `false` | Apply tags only after metadata extraction has finished |
 | `IMMICH_GO_UPLOAD_SKIP_VERIFY_SSL` | `--skip-verify-ssl` | `false` | Skip SSL verification |
 | `IMMICH_GO_UPLOAD_TAG` | `--tag` | `[]` | Add tags to the imported assets. Can be specified multiple times. Hierarchy is supported using a / separator (e.g. 'tag1/subtag1') |
 | `IMMICH_GO_UPLOAD_TIME_ZONE` | `--time-zone` |  | Override the system time zone |

docs/upload-commands-overview.md

@@ -52,6 +52,7 @@ Tags can be assigned in two main ways:
 
 *   **`--folder-as-tags`**: Uses the folder structure to create tags. For example, a file at `Holidays/Summer 2024/photo.jpg` will be tagged `Holidays/Summer 2024`.
 *   **XMP Metadata**: Tags are also read from XMP sidecar files.
+*   **`--defer-tags`**: Postpones tag creation until Immich finishes metadata extraction so built-in file tags/keywords aren’t dropped.
 
 `immich-go` creates new tags on the server as needed and efficiently tags assets in batches.