I’m running a fanless mini-itx msi celeron mobo and haven’t had any issues, except a short living hd (it lasted less than one year, but it has been put to some stress). After the hd crash, the mobo rested forgotten for some years, and now I reactivated it with an ssd.
My point is that this mini server is located in a room without air conditioning, and the summer here in Rio is very hot, so as far as I remember, the cpu temperature never went below 70°C, with an average of almost 80° during the hot period.
Would you be willing to share your spreadsheet system. It sounds great, but I’m not experienced enough with spreadsheets to set up search by time, event, etc.
I realize that I’m replying to a post that is almost 2 years old - but your description of your starting point here is identical to mine - except that I have close to 1,000,000 photos I’ve managed to aggregate to a hard drive.
I’m wondering if you would be willing to share how you finally solved your organizational challenges and any lessons you learned while doing so.
I’ve just bought an Asustor NAS and have a 10TB mirrored RAID array set up with 2 1TB SSD’s for caching and dual 2.5Gbe (aggregated) network connections. The NAS itself has a quad-core Celeron processor and 4GB of RAM. I managed to get DigiKam installed in a Docker container running directly on the NAS device and I can access the DigiKam interface from a web browser on any computer on my network… but I’m completely new to DigiKam.
From a MacBook, I first tried Lightroom Classic to scan external hard drives and build a library. I liked that I could configure Lightroom to COPY each photo it found to another location and create a file structure of yyyy/mm as well as rename the photos… But I think I would rather have had them put in a “/yyyy/yyyy-mm/yyyy-mm-dd” - and since I have to import them again into DigiKam - is it possible to have DigiKam copy and rename the files as they’re imported - or should I use lightroom to do that on another computer and then just copy the resulting structure to the NAS?
Please share any other advice you’d care to offer after having gone through this yourself - and especially anything I need to watch out for.
Oh - one other thing… is there any way (from metadata?) to automatically detect that certain images are screenshots rather than photos - and move them to a separate area? Or that a picture is extremely small/low resolution - so to skip it altogether?
And after I have this wonderful library of images assembled, is there any AI software I can run to examine the photos and automatically create tags for what they find in each photo? (ie Trees, Lakes, Cars, People, banana’s, etc…). Kind of like our iPhones already do with photos in our iPhone library…?
Short answer: I didn’t. I was waiting for Synology Photos which came out middle of 2021 - and it had FAR less functions than the old separate Photo Station and Moments. They want to add functions “later”, but since we have near December 2021 now I believe the time frame will be years.
I may look back into other options (again), Digikam is one. However it seems way over the top and I worry about performance inside the Docker. Another option posted here a short while ago is Photoprism - I like that design more. Seems even more difficult to install. And since it’s new, it has probably a risk of getting stale.
So for now I am sort of stuck with the pathetic Synology Photos. I store my images by year and sometimes some subtopics. However one can simply filter images and tag them. So I am not really concerned about any folder structure.
Now is the Digikam performance on your NAS? From inside and outside your network? Does Digikam has good phone apps? That’s another issue I worry about - to complex phone apps for the average family user.
photoprism (looks nice - see demo - but has essential functions missing, like image rotate/multi image rotate - makes it for me a ‘no no’)
Hey - thanks for the reply - and for the link to PhotoPrism. That’s not one that I’ve come across yet and looks like it’s got some decent features.
DigiKam installed and started up fine in Docker on the NAS. It was fairly responsive in the web browser “guacd” interface - at first. I was able to point it at a photo library on the NAS with about 600,000 photos in it - and it imported them all in just under 24 hours… but then I started having trouble getting the interface to even load in the browser. If I restart the Docker container, it loads for a couple of minutes - but then disappears. I don’t know if it’s trying to generate fingerprints and not leaving enough CPU cycles for the the interface - but if I can’t figure out how to fix it, it’s not going to be viable.
I’m upgrading all of my networking equipment and building some new servers and until I’ve got them all set up, I’ve completely disabled any access to my network from the outside world.
DigiKam talks a good game - with face recognition, duplicate scanning, tags and a plethora of features - but if I can’t get it to run stable, I guess it’s going to be bad to the drawing board. I also think I might be trying to do too much simultaneously and have only been skimming documentation.
The first thing I need to do is aggregate the photos, then combine them into 1 file system structure and remove duplicates. Maybe I should stick with lightroom for the aggregation and organization - and then use a paid tool to find and remove duplicates?
What did you do to scan for and remove your duplicates? There are about 40 programs I’ve come across already and they all claim to be the best. Is there one you’ve used that you would recommend?
I’m back in 2023 and still looking for a way to auto-tag my photos with the stuff that is in them. DigiKam 8.0 was just released and it leverages some deep learning for image quality sorting - but still nothing for tagging.
I’m kind of amazed that nobody has created a plug-in for YOLO or some other PyTorch or Tensorflow image classification. It might also be useful to have some way to detect NSFW images.
It’s frustrating to me that my iPhone has been doing this sort of thing for years - automatically without my asking so that I can type in a search for “eggs” and instantly see all 147 photos that contains eggs - but that it isn’t even an option in DigiKam.
I really hope they are working on a whole host of AI and Machine Learning add-ons - for my classifications needs AND for some cool editing features like Google just announced with their Magic Editor.
Maybe I just need to find a different product. How do I cancel my subscription and get my money back? (kidding) I know it’s a free product - but I think I’d rather pay for a product with the features I need than struggle to use something lesser just because it’s free.
If anyone stumbles onto this thread and can recommend something that lets one search my image library for images containing a bird, a boat, a meal or people wearing hats - please add a comment and tell me! Or maybe just wait another 12 months and by then ChatGPT or Bard should be capable of doing it for me, right?
It is equally frustrating that you expect an application developed by a handful of people, none of them full time or for much in compensation, to keep up with the richest company in the world’s flagship product.
“Why haven’t people worked for free to give me the features I demand?”
Geeeee I don’t know.
I can’t even comment on the rest of your post. Jesus. Fucking. Christ.