Option to have Tropy add tags to actual image files

Would it be possible to have Tropy (optionally) apply tags in the image files’ metadata?
(And other metadata, such as title, creator, etc.)

That would be great for me, because then I could find images by tags for use in other programs. An example of a program that does this is Picasa. It allowed finding things within the program but also on the computer or in other software that could read tags.
Actually, Picasa probably has/had other features that would be useful in Tropy. If only Google had made it open source when they killed it.

As useful, or maybe more useful since I’ve been tagging my images for years, would be for Tropy to be able to recognize and display the metadata already attached to files.

Writing the metadata back to a photos (using XMP most likely) is something we’ve envisioned from the start. I don’t have a schedule for this yet, but it’s definitely something we’d like to do.

As for parsing metadata embedded in photos during import, Tropy is already using some of the metadata (if available), but we’d like to improve that further. Could you post an example photo with embedded metadata and tell me which tags you would expect Tropy to import automatically?

1 Like

The word “automatically” is always scary. Because I’m a pretty heavy user of metadata, I’d want a list of all possible metadata from which I could choose all the fields I need, and the ability to set up import template(s) so I don’t have to choose each time. At a minimum I’d want title, description, author, date taken, rights, tags, comments. And the option to include any data about camera/lens/exposure/GPS/etc. that seems useful, which would vary depending on the image’s origin and purpose.

Here’s an image that has most of the fields I’d want (download link is below the image):

(I don’t know if the one shown here still has its metadata, but the linked one should.)
I could also export an .xmp file from Photoshop if that would be useful.
Or you can see a lot of the EXIF data here: Juvenile Shortnose Gar (Lepisosteus platostomus) from nort… | Flickr

It’s unfortunate that not all file types allow tagging/metadata. I’ve been getting images of portions of newspaper pages from Chronicling America (by zooming in to what I want and then right-clicking and saving image–more efficient than downloading a PDF or JP2 of the whole page and cropping it), but they’re PNG files. Can’t be tagged, so I guess I’ll have to batch convert to JPG if I want to work with their metadata (or want Tropy to eventually be able to write metadata to them).

1 Like

In this example the metadata is XMP so most of it should already be extracted by Tropy (never mind that the metadata panel currently does not treat longer text very well). From what I can tell, we’re missing the dc:subject field (your tags!) but you should see title, creator, description, and rights filled in, right?

The reason why we’re not importing dc:subject yet is actually because I’ve been wondering if we should create tags from that field automatically (in your case, I think this would be the right thing to do?).

I chose that image because it had tags, title, description, etc., but it wasn’t one I’d added to Tropy. Oops. It was literally the first image I found on my computer that had those elements. I have since added it, and am happy to see that Tropy picked up some metadata. Tags are still the important one for me, but this is good news for sure.

It’s not hard to copy and paste from the description (or title etc) field into the notes area of Tropy, but a right-click context menu item saying “Copy to notes” might be useful. For any/all metadata fields, I think.

May I ask, is XMP the requisite metadata connection-space, so to speak? Or would it be possible to use xattr dictionaries to add metadata to formats on a case-by-case basis?

Like, when converting from a pdf or png to a jpeg, the fallback for any fields in the source format metadata without analogues in the destination format would be to store them in an xattr field with prefixed key names (e.g. something like com.tropy.md.PNG.chunk.sTER) – that way the metadata is rather efficiently stored “in” the file, instead of an XMP sidecar.

That’d be neat to have, as like a general metadata fallback option; tbh I don’t know if whether that would fix these problems, or just create more of them – if the latter, I’d be into it. Yes!

Tropy is currently extracting XMP and EXIF metadata on import (only properties defined by your template are used). Eventually, we’d like Tropy to be able to write your current metadata back into the original files (in all likelihood this is going to be a separate action, something like ‘write metadata to file’), we will almost certainly use XMP to embed the metadata directly in the file (depending on file type an XMP sidecar may be required, but most image formats support embedding XMP). (We don’t have a timeline for this at the moment.)

1 Like

I am very interested in this possibility to write the metadata to files, even if it is limited to a small group of IPTCs. At the moment I am still using Adobe’s tools (Bridge) to do that. Do you have a development schedule for this feature now? Thanks a lot.

This is not yet on our current development cycle, sorry.