For some time now I’ve become more and more interested in running my own services and using less and less external services. The latest has been Flickr, which I had been a Pro member of for over 8 years now (yikes!). I used it less and less, and was more wary of uploading pictures to it, so I thought it made sense to just paying for it. However, there was one thing I was still using Flickr for, a program I wrote called clj-flickr-memories (now clj-photo-memories): every week, that program searches for pictures taken this-week-several-years-ago and sends you an e-mail with the results, if any.
The first possible alternative I found was MediaGoblin, but the API is too basic and it doesn’t even save the picture details (such as the date the photo was taken) in the database so I couldn’t even improve the API to support that. I was close to giving up when I found Trovebox: it’s a hosted service you can pay for, like Flickr, but it’s also open source so you can host it yourself if you want. And although its API documentation leaves a bit to be desired, it could do what I wanted, so I got cracking and modified my clj-flickr-memories to support both Flickr and Trovebox.
If you want to switch from Flickr to a self-hosted Trovebox and want to import your photos, there are two things you should know:
If you self-host and use HTTPS (and you should!) you need to include the “https://” both in the “host” key in the command-line tool configuration file and in the Android app.
You can easily import all your Flickr photos in two steps: first you use export-flickr to fetch information about your Flickr photos (only works with Flickr Pro accounts, though!) and second you use import to import those photos into Trovebox. Note that the first step will leave a directory “fetched/” with one file per photo, so you can choose which photos to import to Trovebox, eg. based on the date.
Trovebox also has mobile apps, but at least the Android one requires internet access to it’s not great for me (you can’t browse photos offline).