<div dir="ltr">Hi,<div><br></div><div>I tried to use DK with a remote server (LAN) and storing the thumbnails locally, but for me it was too slow so I went back to keep all DB´s locally. So unless you are using a cabled network remote server is not a good solution in my opinion.</div><div>In addition to storing the photos on a NAS I backup to Google Photos (GP). Being clever I tried to store the faces into the image metadata so that any computer could recognize the faces. If it hadn´t been for GP reading the changed photos as new photos that would have worked fine I guess. When GP saw the "new" photos, Google Drive started to back them up again and created a lot of duplicates in GP. Photos that were taken in 2005 suddenly appeared with today's date. Luckily all of them had the same modified date so I was able to delete them all quite easily.</div><div>If GP could read "date taken" instead of "modified date" I would probably have chosen to store faces in the metadata.</div><div><br></div><div>dajomu</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Feb 3, 2022 at 2:56 PM Henrik Hemrin <<a href="mailto:hehemrin@hemrin.com">hehemrin@hemrin.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="overflow-wrap: break-word;"><div>Hi,</div><div><br></div><div>Not the answer but factual background, taken from The digiKam Handbook, page 30:</div><div><br></div><div>"Everyone knows about database, it is used to store data. As all other photographs management programs, digiKam too uses the database for some obvious reasons like avoiding data duplica- tion, reducing data redundancy, a quick search engine, and greater data integrity. Moreover, the cost of data entry, storage and retrieval are drastically reduced. Additionally, any user can access the data using query language.<br><br>Talking in particular about digiKam, the Albums, Album Roots, Tags, Thumbnails, Face Recog- nition Data, Image Metadata, File Paths, Settings etc. are all stored in different database files.</div><div><br></div><div>The digiKam actually manages more than one databases. For convenience, it is broadly categorized in three:<br><br><div><span style="white-space:pre-wrap"> </span>• Core database for all collection properties, i.e. it hosts all albums, images and searches data.<br><br></div><div><span style="white-space:pre-wrap"> </span>• Thumbnails database for compressed thumbnails i.e. to host image thumbs using wavelets compression images (PGF format).<br><br></div><div><span style="white-space:pre-wrap"> </span>• Face database for storing face recognition metadata i.e. to host face histograms for faces recognition.”</div></div><div><br></div><div>Regards</div><div>Henrik</div><div><br></div><br><div><br><blockquote type="cite"><div>3 feb. 2022 kl. 14:09 skrev George Koulomzin <<a href="mailto:george@koulomzin.com" target="_blank">george@koulomzin.com</a>>:</div><br><div>
<div><p>Thanks for your reply!</p><p>The XMP solution would avoid issues with invalid paths that I
suspect would be a problem if one were to try to copy the actual
DB, so it is attractive.</p><p>I've also been reading up on the "remote server" approach. I
would be the best solution if performance is acceptable. Does
anyone have any experience with this? <br>
</p><p>It is not clear to me where the "training" information is kept,
i.e. the set of parameters which help the algorithm decide this
picture shows John rather than Joseph. Clearly, sidecars would
not be the place. How would one synchronize <i>that</i>
information?<br>
</p><p>My environment is:</p>
<ul>
<li>Windows 10 OS all around.<br>
</li>
<li>WiFi everywhere, with cat 6 cable to some locations</li>
<li>90,000+ image files, with a lot of "duplication" -- I keep
both jpg's and raw files.</li>
</ul><p>Thanks in advance...</p><p><br>
</p>
<div>On 2/2/2022 7:27 AM, Henrik Hemrin
wrote:<br>
</div>
<blockquote type="cite">
George,
<div><br>
</div>
<div>On the second question, I would use the XMP sidecar
as the key for solution. You set up so DK writes metadata to XMP
side car, you bring the latest side car with you between your
computers and ask DK data base to read the latest XMP side car
file after you have updated it/brought to the other computer.
This method of course means you cannot update the same photo on
both machines at the same time. </div>
<div><br>
</div>
<div>Someone else may elaborate more in detail, but one
general level, I believe this can be a simple solution. </div>
<div><br>
</div>
<div>I am sure it is possible make a more advanced setup,
I guess that kind of effort depends on how many photos, how
often etc, if a more automatic proper sync is better. But again,
someone else may have that answer. </div>
<div><br>
</div>
<div>Best regards</div>
<div>Henrik Hemrin<br>
<div><br>
<blockquote type="cite">
<div>1 feb. 2022 kl. 14:25 skrev George Koulomzin
<<a href="mailto:george@koulomzin.com" target="_blank">george@koulomzin.com</a>>:</div>
<br>
<div>
<div><p>Hello, I a new to the list, so please
forgive me if these questions have been answered.</p>
<ul>
<li>the archive is organized by month. Is
there a way of seeing thread topics by year, or
longer?</li>
<li>how can one synchronize tag metadata
(e.g. faces) across two computers. I have a desktop
pc with ALL my photos and a laptop with SOME of my
photos (laptop is always a subset). I've done face
recognition on the laptop (subset); can I port the
resulting metadata back to the pc so as not to have
to repeat all the training?</li>
</ul><p>Thanks in advance!<br>
</p>
<pre cols="72">--
George Koulomzin
7 Bridge Hill Lane,
PO Box 781
Bridgehampton, NY, 11932
(h) 631-537-4956
(c) 914-393-6179</pre>
</div>
</div>
</blockquote>
</div>
<br>
</div>
</blockquote>
<pre cols="72">--
George Koulomzin
7 Bridge Hill Lane,
PO Box 781
Bridgehampton, NY, 11932
(h) 631-537-4956
(c) 914-393-6179</pre>
</div>
</div></blockquote></div><br></div></blockquote></div>