Migrate: Read metadata from Apple iPhoto / Photos app XMP

[ylluminate]

I have a rather large photo and video library that spans a few decades and is over 2TB in size. Up to now I've been managing it with Apple's ecosystem, but the time has come to move away from them for various reasons.

Is there a sane migration path that exists to facilitate taking all of the contents of an iPhoto / Photos.app library over into a PhotoPrism setup?

---

Acceptance Criteria:

• Read title from Apples XMP
• Read keywords from Apples XMP
• Read takenAt from Apples XMP
• Read GPS information from Apples XMP

[lastzero]

Excellent question, didn't test this yet - only with Google Photos, which provides JSON files in its export. Wondering how Apple stores metadata? Maybe inside image files. In that case, we might be able to just index it.

[swingstate]

I am working through the same hassle with a 1TB library, 99% IOS based photos and videos.
I am testing imports through the web-based import and even with a chunk of 4000 files at once this worked well. Import of HEIC photos works very well.
However, photoprism does not seem to support the video codecs which Apple uses to shot. Hence, I am not sure if I'll end with this solution. Going through another step to convert videos first and then upload doesn't make sense to me resp. I want to automate this.
Be aware, the WebUI doesn't allow to delete photos (I don't get why as this is a basic file management task since Windows 3.x) However, deleting files seem to work through WebDAV which I still need to test.

[lastzero]

Any reason you import / upload the files? Indexing the iPhoto library should be easier / faster?

Deleting was not implemented as many users didn't want us to change anything in their library at all, that's also why we added the read-only mode. Guess we'll implement this now together with a config option to disabled it. Note that there already is a delete button for sidecar files.

Transcoding is also on our todo. Financial support enables us to spend more time with this project and develop features faster.

[swingstate]

To your last point, I am considering to donate, however want to be sure where and for what. For me it's basically a long-term decision to move away from the vendor lock-in to a self-hosted library. I like the archive feature to move unwanted files to one place, However, there might be shots I don't need anymore (often a lot) and I prefer to keep my lib as clean as possible.

Regarding the video transcoding - I am testing handbrake atm, which eventually could help for the interim to convert from 265 to 264 and to put the output in the photoprism import folder.

Regarding first point - My workflow in this self-hosted setup will probably very simple.

1. I shoot with the Iphone and use an app to automatically upload via SSH connection to an folder on my UnRaid host.
2. I browse the folder with a Webbased filebrowser to scan through the photos and remove the stuff I don't want in my library
3. I move the photos which I want to keep to the photoprism import folder (didn't test that feature yet, hope it exists)
4. I sort the imported photos into albums

Event though this flow is simple it forces me to decide on the photos to keep before I move them into Photoprism. Going there later via WebDav to delete photos is only a workaround and so far at leat my Win10 host struggles to connect. Didnt test it on my Mac yet.

So my point is simplicity. Let me dump all my files into one import folder in the file system and then let me decide later in the photoprism app where I want to put photos and which I want to keep.
I think I get your point regarding read only and the UI is simple and clean. I don't aim to destroy that, hence if you could enable the delete function as an "to be activated" feature you would earn one more lasting customer. Video transcoding can be added later I think.

[lastzero]

Thanks for your detailed feedback! Helps a lot. Funding details are documented on https://docs.photoprism.org/funding/.

[Engr-AllanG]

I agree an option to delete would be ideal. My workflow is to use photoprism as a one stop shop - after initially importing my library (mostly complete) I will be auto uploading from my phone (primary source of photos) with my seafile server using the android/ios auto upload feature (and then webdav hourly sync between photoprism and seafile). My point is that I will have a lot of junk photos I will need to delete. Storage is cheap so archiving isn't terrible, but would prefer to remove permanently

I too will be donating. I contribute a small monthly amount to another application I have found very useful.

[swingstate]

Before moving my entire library into the app - wanted to confirm whether the delete function is coming. Or is this still being reviewed?

[lastzero]

@swingstate We're going to implement #167, just can't tell you when exactly as there are ton of feature requests we need to work through (like face recognition and batch edit).

[swingstate]

ok, cool! Didn't mean to spam as I am still learning how GitHub works. I'll post a few more ideas shouldn't there be any existing entries.

[swingstate]

I'll give it a try and meanwhile imported the first chunk of 1500 Photos. Any chance I can map the Archive through WebDAV to delete unwanted photos? Or can I somehow list the files (filenames) which are in the archive? Even in the terminal would be fine.

[lastzero]

Our WebDAV implementation is based on the actual file system, not database contents. Guess we really need to implement #167 soon...

[shoeper]

@swingstate just a quick note regarding video conversion: When you have a graphics card it could be much faster to utilize it which is possible e.g. using ffmpeg. I recently tried several methods and using an Nvidia (GTX 1060) graphics card was by far the fastest solution (and also seemed to result in better quality compared with software encoding).

[swingstate]

I set up following .mov conversion flow: Upload via WebDav into Nextcloud -> Conversion with Handbrake Docker container -> import into photoprism.

…

Am 02.12.2020 um 19:58 schrieb Sven ***@***.***>: @swingstate <https://github.com/swingstate> just a quick note regarding video conversion: When you have a graphics card it could be much faster to utilize it which is possible e.g. using ffmpeg. I recently tried several methods and using an Nvidia (GTX 1060) graphics card was by far the fastest solution (and also seemed to result in better quality compared with software encoding). — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#636 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ARSCO2V22CMDUFOQ4R55NGDSS2E6VANCNFSM4T6YNSEQ>.