Just installed Gramps 5.1.4 for Windows, though I have been using it off and on for quite some time and am still subscribed to the old sourceforge mailing list. A recent announcement about the developer for gexiv2 wishing to step down caught my eye. Not that I want to step in as a replacement - my experience with Gtk & github, Python etc is way too limited to take on something like that, but …
It got me curious about the plugin mentioned Editexifmetadata , which seems to have fallen by the wayside.
There seems to be limited display of media metadata in 5.1.4, presumably via gexiv2, and I am curious if there is any way to customize or add to the data displayed in the metadata tab.
In particular, I am more interested in XMP or IPTC data describing the content of the image, rather than merely data about the size & data of the image.
As part of my posting here I have found an earlier thread and will have to have a closer look at the thread later, since that thread is closed.
I am not so much interested in editing, though it might be nice to fix typos and other such ‘small issues’, but it would be very helpful if one could at least review the data.
As well, I am curious if anyone has suggestions as to which tags are useful or which to avoid.
I much appreciate the feedback;
Unfortunately I am not familiar enough with experimenting with patches to be able to try and check out this feature.
Still I am hoping that the code will eventually make it into the mainstream.
Any estimates on when that might happen?
Almost all of my experience is in C++ on Windows and I don’t have enough background with Python, specifically with debugging anything in the rather large Gramps universe.
At one time I had been working with and building Gramps AIO packages for Windows, but never with the Python code itself.
Since then, I have been sidetracked into trying to build a utility which would make it possible to add enough metadata to images etc to make it useful in any case.
Making it usable and useful under Gramps is a more remote goal on that track.
Fortunately, installing patches is fairly straightforward… even for those of us not using any of the GitHub automation.
The top-level of the patch tell you how many files have had changes, which files, their filepaths (in relation to the Gramps application installed location) & highlights the differences. (Those differences offer a good introductory approach to understanding Gramps & Python.)
The first impression (that you will have go through manually and edit each changed line) is alarming but erroneous. It is easier than that.
It is definitely worthwhile starting Gramps from the command line on occasion. (And certainly when installing something new.) Because all the complaints and grumblies that are usually invisible become very accessible.
The warnings are piped to the command line console and remain visible even if Gramps spontaneously closes itself.
Would you share your progress with us occasionally? I recently did an experiment with Picasa and intended to bring in the metadata to Gramps. The machine learning for face recognition was very interesting but it was all XMP that not even Google’s cloud-based replacement tool recognized.
I am wondering if I brought in samples from Picasa and Google Images file, whether it would be possible to extend the a file’s metadata so that it would be readable by either?
One of my big questions is how to make such data portable for genealogy? For instance, maybe there is a focal person for a group and each person might have a genealogical numbering system reference number relative to that person?
So, for a family reunion photo, the metadata might include the focal ancestor (who is generations past & not even in the photo) and metadata includes a Modified Henry Numbering system (with Spouse S1, S2, S3… suffix where necessary) to facilitate associating the persons if the photo is imported into a foreign tree. Once you crosswalk the reference person, there’d be some intelligence for refining matches on the others.
I have no experience (yet) with adding face recognition data to metadata, though I have used Picasa in the past, before Google abandoned it.
From what I know now, there is a fair bit of face recognition related information support in metadata, though I have not done anything with it nor really checked out any images with such data.
On the whole genealogy & metadata covers a lot and I am surprised there is not much I have been able to find regarding its usage for genealogy…
In fact even inspecting/editing and adding information to metadata is not well served, IMO.
There are several areas, in this respect which need to be considered (and worked on)
as mentioned, just being able to inspect … metadata for images and other related documents, mainly PDF
figuring out which metadata is adaptable to genealogy - up to now, the metadata ‘bus’ was/is driven mainly by news, image & data processing companies; recently there seems to be some development in the genealogical community to get involved in using metadata: https://fhmwg.wordpress.com/
adapting genealogical software to at least recognize and display some metadata, and eventually more and more genealogically related data
adapting genealogical software to possibly analyze & use some of this metadata
As for sharing my progress, this is precisely why I started this thread.
As my own project to at least inspect and modify image metadata for JPG evolves and become usable, finding answers to the rest of the issues becomes more important.
It is because I wanted to embed much of this information into my images that got me started on investigating metadata.
The software to inspect or modify the data was mainly command line based, which is a bit tedious for most of us. There is some GUI related software, but none seemed to do what I thought I needed/wanted.
My current app for metadata inspection is getting close to workable, with fewer and fewer warts and holes. Once it is ready for beta testing I most certainly want to make it available, though for a start it will likely remain closed source. It is/was inspired by (the now abandoned) ExiftoolGUI from Bogdan Hrastnik : https://exiftool.org/gui/
As for your question regarding those Henry numbers I have no answers myself, but I’d encourage you and anyone else interesting using metadata for genealogy work to investigate the working group and become involved: https://fhmwg.wordpress.com/
FWIW, I have just found ImageRanger - not free, but reasonable & not cloud-based - and it allows me to add face data which I can read back via Exiftool.
It would seem that if one wanted to add Henry numbers, the simplest place (and IMO, most logical place, unless we can find a specific location for such data) would be as part of the name associated with the face.
I have Digikam, but was not able to find out how to add face data nor how to make sure it is written to the image.
I do admin to not spending a lot of time to find out how, but I am pleased to know that DK can do so and will investigate that facility some more - as time permits
Any good places to find a howto?
Suppose I have a family picture of some ancestors.
What I need, is to find out how to use Digikam to have it detect the faces and how to add names to go with each face.
Then I need a way to ensure the data is written to the metadata in the image.
After that, I am sure I can read back any data which has been written.
There seem to be at least 2 ‘formats’, i.e. groups of metadata tags which are commonly used for this purpose.
I am not really familiar with any of these formats, but since I can read back all there is, I would expect that with a few samples, we can sort out what goes where.
Somewhere along the way, I would also expect the FHMWG will eventually - don’t recall right now whether they have done so - define preferred or recommended metadata tag groups.
If they do, they most likely will also provide some sample code to inspect the fields they recommend.
I have just released a utility to read and modify image metadata for JPG, PNG & TIF files.
It is simply a GUI front-end for the Exiftool library and currently only runs under Windows. As well, it is intended to support multiple languages, though that part is currently very much in its ‘infancy’ because my experience in that area is very recent and limited.
It was inspired by the no-longer-supported ExiftoolGUI app, though there are a number of differences.
The project was motivated by my need to add metadata directly to the image files for my genealogical work, rather than to rely on file naming and directory tree structure alone.
Yes, I did look into named regions etc, though I more or less forgot about the posting.
And yes, the extra data would be of interest and I have to start looking at that patch. Still, all of that is a bit scary right now, because I have never really worked much with patching etc, least of all github. Still, I do use git for my local code management. so it is just a matter of getting used to some new things.
I just finished getting the latest Gramps code from Github to run on my VM, but will have to also look into installing the Edit Exif Metadata plugin to see where it is at.
Finally did get a chance to follow up on this.
FWIW, I have given up on my VM-base Gramps project and instead installed Gramps on a stand-alone Mint box.
But first I had to try and find which of the several copies of Gramps hanging around on my systems had the Metadata patch installed.
(Side question: how do the developers keep track of things like which copy has what?
No doubt, when one works on such a project more seriously, one might not need such
crutches, but for myself, where it might be weeks between getting a chance to follow up,
some details just disappear in the dust & the ‘rearview mirror’.)
Back to the topic:
my first impression is that it works, though I have not yet seen any thumbnails for any of the images I have tried.
as it seems that the purpose of the gramplet is to display metadata from other sections than EXIF, perhaps ‘Exif’ should be dropped from the gramplet name.
much of the displayed metadata, especially several fields from EXIF, IMO, are of little, if any relevance for genealogy, such as Pixel?Dimension and could easily be dropped to make more room for the more important fields from XMP or IPTC specs.
From what I have seen, this version looks very promising and I will be working with it from here on out.