Update photoprism to 260305-fad9d5395 (#5206)

photoprism/docker-compose.yml

@@ -10,7 +10,7 @@ services:
       PROXY_AUTH_WHITELIST: "/originals/*,/import/*"
   
   web:
-    image: photoprism/photoprism:251130@sha256:db16ee6b1ba3e5bc5ad787a3141a082532d1b30932c8b1396fa609394374b83f
+    image: photoprism/photoprism:260305@sha256:5229f68bce0293320547f35c733de584d7c2d5725a0efbddb68d7a118a8fc21c
     user: "1000:1000"
     restart: on-failure
     stop_grace_period: "1m"

photoprism/umbrel-app.yml

@@ -2,7 +2,7 @@ manifestVersion: 1
 id: photoprism
 category: files
 name: PhotoPrism
-version: "251130-b3068414c"
+version: "260305-fad9d5395"
 tagline: Self-host your photo and video library
 description: >-
   PhotoPrism® is a privately hosted app for browsing, organizing, and
@@ -42,23 +42,23 @@ defaultUsername: "admin"
 deterministicPassword: true
 torOnly: false
 releaseNotes: >-
-  This major update introduces the long-awaited Batch Edit dialog for editing metadata of multiple pictures at once, along with an upgraded face recognition pipeline that delivers more and better matches. PhotoPrism now integrates directly with Ollama and OpenAI to generate captions and labels, with support for custom models and refined configuration options.
+  This service release focuses on security hardening, interoperability improvements, and bug fixes.
 
 
   Key features and improvements:
-    - Added Batch Edit dialog to edit multiple pictures at once
-    - Improved face detection and recognition with a new engine
-    - Direct Ollama integration for generating captions and labels
-    - OpenAI integration for generating captions and labels
-    - Support for custom image classification models
-    - Added menu to select a cover image for people
-    - Add pictures to multiple albums in one step
-    - Improved people name editing and focus management
-    - Enhanced window scroll position restoration when navigating
-    - Added scheduling options for running AI models in the background
-    - Improved memory and event management in Viewer and Places
-    - Fixed issues with video transcoding and playback
-    - Enhanced metadata handling and error logging
+    - Added support for Ollama configuration via OLLAMA_BASE_URL and OLLAMA_API_KEY environment variables
+    - Improved Ollama caption generation with better fallback handling for reasoning models
+    - Fixed merged photos incorrectly keeping image type after video file merges
+    - Fixed GPS coordinate handling near map boundaries in Places
+    - Fixed emoji subfolders conflicting with parent folder albums
+    - Fixed folder album path collisions caused by truncated slugs
+    - Hidden results now display file error reasons in Card and List views
+    - Improved WebDAV response headers for better interoperability
+    - Added HTTP security hardening configuration options
+    - Improved Gzip route exclusion performance
+    - Fixed issues when hosting on a shared domain
+    - Improved compatibility with Google OIDC identity provider
+    - Updated French, German, Latvian, and Romanian translations
 
 
   Full release notes can be found at: https://github.com/photoprism/photoprism/releases/