[TASK] Delete assets that are deleted in Immich but exist on the local disk

.env.example

@@ -2,3 +2,6 @@ IMMICH_HOST=http://localhost:2283/api
 IMMICH_API_KEY=
 
 TARGET_DIR=./export
+
+# File extensions passed in this array will be deleted by the script if they're no longer found in immich
+# MANAGED_FILE_TYPES=

README.md

@@ -5,7 +5,7 @@ Script to export immich data as a structured folder. Useful if you want to brows
 ## To-do
 
 - [x] Create folders like "Misc {{year}}" for photos that are not in albums
-- [ ] Move (or rename?) photos that are deleted in Immich but exist on the local disk
+- [x] Delete assets that are deleted in Immich but exist on the local disk
 - [x] Dockerize the tool so it's easy to run on the server
 - [ ] Stretch: Preserve metadata (e.g. file creation times)
 - [ ] Stretch: directly sync to Proton Drive (skipping the local filesystem)

immich-export.py

@@ -1,7 +1,8 @@
 import os
 import re
 import argparse
-from dotenv import load_dotenv
+
+import dotenv
 
 import generated.immich.openapi_client as openapi_client
 from generated.immich.openapi_client.rest import ApiException
@@ -11,13 +12,18 @@ parser = argparse.ArgumentParser(description='Export albums from Immich to local
 parser.add_argument('--dry-run', action='store_true', help='Print actions without writing files')
 args = parser.parse_args()
 
-load_dotenv()
+dotenv.load_dotenv()
 
 configuration = openapi_client.Configuration(
     host = os.environ['IMMICH_HOST'],
 	api_key = { 'api_key': os.environ['IMMICH_API_KEY'] },
 )
 
+managed_file_type_str = os.environ.get('MANAGED_FILE_TYPES', 'jpg,jpeg,png,gif,bmp,webp,tif,tiff,wmv,mp4,mkv,dng,flv,mpg,3gp,mov,avi,webm,m4v,wmv')
+managed_file_types = set(managed_file_type_str.split(','))
+
+delete_threshold = os.environ.get('DELETE_THRESHOLD', 0.2)
+
 # Set target directory from env var or use default
 target_dir = os.getenv('TARGET_DIR', './export')
 if not os.path.exists(target_dir) and not args.dry_run:
@@ -42,11 +48,16 @@ def save_asset(asset, asset_dir):
     asset_filename = asset.original_file_name if asset.original_file_name else f"{asset.id}.jpg"
     sanitized_asset_filename = sanitize_filename(asset_filename)
     asset_path = os.path.join(asset_dir, sanitized_asset_filename)
-    count = 1
+    if asset_path in asset_paths:
-    while asset_path in asset_paths:
+        asset_filename_without_extension = os.path.splitext(sanitized_asset_filename)[0]
-        # Duplicate filename, add a number to the end
+        asset_extension = os.path.splitext(sanitized_asset_filename)[1]
-        asset_path = os.path.join(asset_dir, f"{sanitized_asset_filename} ({count})")
+        count = 1
-        count += 1
+        while asset_path in asset_paths:
+            # Duplicate filename, add a number to the end
+            asset_path = os.path.join(asset_dir, f"{asset_filename_without_extension} ({count}){asset_extension}")
+            count += 1
+
+    asset_paths.add(asset_path)
 
     # Skip if file already exists
     if os.path.exists(asset_path):
@@ -66,7 +77,7 @@ def save_asset(asset, asset_dir):
     return asset_path
 
 album_count = 0
-copied_assets = set()
+copied_asset_ids = set()
 asset_paths = set()
 
 # Enter a context with an instance of the API client
@@ -107,7 +118,7 @@ with openapi_client.ApiClient(configuration) as immich_client:
                         # Add a small delay to avoid overwhelming the server
                         # time.sleep(0.5)
 
-                        copied_assets.add(asset.id)
+                        copied_asset_ids.add(asset.id)
 
                     except ApiException as e:
                         print(f"  Error downloading asset {asset.id}: {e}")
@@ -129,7 +140,7 @@ with openapi_client.ApiClient(configuration) as immich_client:
         assets_not_in_albums = search_api.search_assets(search_query)
         while assets_not_in_albums:
             for asset in assets_not_in_albums.assets.items:
-                if asset.id in copied_assets:
+                if asset.id in copied_asset_ids:
                     continue
                 # Create a folder for the asset's year (i.e. Misc {{year}}) if it does not exist yet
                 year_folder_path = os.path.join(target_dir, f"Misc {asset.local_date_time.year}") if asset.local_date_time else "Misc"
@@ -139,7 +150,7 @@ with openapi_client.ApiClient(configuration) as immich_client:
                         print(f"Created directory for asset: {asset.id}")
 
                 save_asset(asset, year_folder_path)
-                copied_assets.add(asset.id)
+                copied_asset_ids.add(asset.id)
 
             if assets_not_in_albums.assets.next_page is not None:
                 search_query.page = int(assets_not_in_albums.assets.next_page)
@@ -152,6 +163,30 @@ with openapi_client.ApiClient(configuration) as immich_client:
     except Exception as e:
         print(f"Unexpected error: {e}")
 
+all_extensions = set(map(lambda x: os.path.splitext(x)[1].lower(), asset_paths))
+print(f"Found extensions: {list(all_extensions)}")
 
+# Now we check for files in the export directory that we no longer found in Immich, and delete those
+to_delete = []
+for root, dirs, files in os.walk(target_dir):
+    for file in files:
+        # Check if the file is in one of the whitelisted file types
+        extension = os.path.splitext(file)[1].lower()[1:]
+        if not extension in managed_file_types:
+            continue
+        file_path = os.path.join(root, file)
+        if file_path not in asset_paths:
+            to_delete.append(file_path)
 
-print(f"Finished processing {album_count} albums with {len(copied_assets)} assets")
+# Note that if more than delete_threshold times the total number of assets are deleted, we won't delete any files, as there is probably a bug somewhere
+if len(to_delete) > delete_threshold * len(asset_paths):
+    raise Exception(f"Found {len(to_delete)} files that are no longer in Immich. This seems to be a large number. No files will be deleted.")
+else:
+    for file_path in to_delete:
+        if not args.dry_run:
+            os.remove(file_path)
+            print(f"Deleted file: {file_path}")
+        else:
+            print(f"Would delete file: {file_path}")
+
+print(f"Finished processing {album_count} albums with {len(copied_asset_ids)} assets")