1. Open the 📷 Memories app in Nextcloud and set the directory containing your photos. Photos from this directory will be displayed in the timeline, including any photos in nested subdirectories.
The exact steps depend on your Nextcloud platform. If you use Docker for your Nextcloud instance, you can install Exiftool by using a custom docker image.
1. ☁ Clone this into your `apps` folder of your Nextcloud.
1. 👩💻 In a terminal, run the command `make dev-setup` to install the dependencies.
1. 🏗 Then to build the Javascript whenever you make changes, run `make build-js`. To create a pull request use `make build-js-production`. Watch changes with: `make watch-js`.
1. ✅ Enable the app through the app management of your Nextcloud.
The approach of this app is fundamentally different from the official Nextcloud Photos app, which is very lightweight and works entirely using webdav. This app instead maintains special metadata in a separate table on the backend, and thus can be considered to have different objectives.
- Once properly configured, Memories is **extremely fast**, possibly one of the fastest web photo viewers.
- On a server with relatively cheap hardware (`Intel Pentium G6400 / 8GB RAM / SSD`), loading the timeline takes only `~400ms` without cache on a laptop (`Intel Core i5-1035G1 / Windows 11 / Chrome`) for a library of `~17000 photos` totaling `100GB`. The test was performed on Nextcloud 24 with `nginx`, `php-fpm` and `mariadb` running in Docker.
- For best performance, install the [preview generator](https://github.com/rullzer/previewgenerator) and make sure HTTP/2 is enabled for your Nextcloud instance.
- You may need to configure the Nextcloud preview generator and Imagemagick / ffmpeg to support all types of images and videos (e.g. HEIC). If using the official docker image, add `OC\Preview\HEIC` to `enabledPreviewProviders` in your `config.php`.
- If local time is not found in the photo (especially for videos), the server timezone is used.
- Indexing may be slow, since all files must be downloaded from the storage. The app currently assumes that the Exif data is present with the first 20MB of each file.