I did a photobook workflow without Google and I’m delighted.

So there’s this family tradition of mine that every year, since we’ve had kids, I’m making a photo book of the highlights of said year. The original, slightly apocalyptic idea was actually to have something physical that we can look at, sitting around the fire, when all of this (waves hands around) is gone and collapsed. (And we have firewood and don’t have to burn our books to not freeze to death.) But even without the Mad Maxian take, it’s a nice tradition, makes a nice Christmas present to everyone, and gives me a couple of sleepless nights, when I realize we are almost on deadline and sit and collect photos and make the thing in about 2 bursts.

Since forever, I heavily relied on Google Photos for this excercise, but this year I have built my little workflow on Nextcloud (and Linux), and I saw that it’s more convenient and quicker. So read on and I’ll tell you more.

The workflow in Google Photos was supposed to be simple: all my photos synced from the phones, cameras, etc. to Google Photos, and around beginning December I went through last year’s harvest, multi-selected the ~200 photo that made up the first filter, and added them to an Album. My wife did the same with her stuff, shared the album with me, and we have our first selection that would be the basis of the book edit.

Difficulty enters here though.

1. First off, we are talking about a book; photos need to be edited, corrected for contrast, histogram, etc… Google’s editor is not bad per se, but not a very usable one when it comes to editing massive numbers of photos. For those I need my GIMP setup, with my hotkeys, my scripts.
2. Editing my wife’s shared album also doesn’t work, because of the sharing/access model of Google Photos.
3. And the multi-select in Google Photos is also becoming very slow and cumbersome with large numbers of photos.

Don’t get me wrong, Google Photos is good. Very good. But it’s a cloud product, and as such will never be as fast as a desktop setup with local files.

So at the end of the day, for multiple reasons, I always downloaded the original filter (those 200+ files), and continued the process on my own laptop. But even with this process, the slowness of the web based photo gallery and multi-select impacted me badly.

Enter Nextcloud.

I’ve been playing with moving to Nextcloud for photo storage on and off for a while, and eventually since Jan 2020 I’ve made Nextcloud my primary photo backup/storage: the Nextcloud iOS app takes care about automatic sync, and although I had some reservations about performance, overall Nextcloud Photos is more featureful for what’s important to me (mapping, etc.), and for the time being I’ve worked around the performance problem with an unencrypted account created just for photos, so that the preview generator plugin can do its job and generate thumbnails periodically. (And yes, this means my photos are not encrypted, but neither is Google Photos to be honest. At least here I know and trust the sysadmin — mostly…)

The other thing of course made possible with Nextcloud is proper desktop sync: the desktop app supports multiple accounts (duh), so I could simply set my Nextcloud Desktop app to sync my Photos folder from that second account. Behold, local copies of my photos, meaning I can use a local photo viewer/editor to speed up my workflow.

Enter Shotwell.

sudo apt install shotwell

Shotwell (I think) is not part of the default Mint setup, but installing it is of course is trivial. Once installed, it imports all photos it finds in the Photos folder. It does a whole lot of things I’m not going to cover, and it does 2 things that helped me tremendously with my photobook project: saved searches and tagging.

My workflow basically got very simple:

1. I ceated a Saved Search called Photobook 2021 range, setting it to start on the day I finalised last year’s book and the day I finalise the 2021 book;
2. I went through this Saved Search and tagged all photos that made the first selection “Photobook”;
3. once I was done, I selected all photos with the “Photobook” tag, and exported them into my “Photobook raw” folder.

I do this third bit so that I can modify the files freely enabling a smoother edit workflow for the photobook software:

# setting file modification dates to exif modify dates wherever I can:
exiv2 -T rename * 

# enabling work on files with space in their filenames:
IFS=$‘’ 

# and finally, prefixing all filenames with the date taken so that I can 
# simply sort by file name regardless of whether a photo was taken with
# an iPhone, a scanner, or my Olympus MILC:
for i in *;do
 mv $i `ls --time-style=long-iso -l "$i"|awk '{ print $6 }'`__$i;
done

Well there.

Not a very complicated workflow, but as I said, hugely speeded (sped?) up by switching to local file processing, enabled by switching to Nextcloud for file backup/storage.

Header image: one of my photos of Kennemerland National Park that made it into the 2021 book.