[digiKam-users] beta 7.2.0 similarity search

[Mark Mara]

I am currently using digikam 7.1.0, (default database(s). It meets or exceeds my needs in all areas except performance. I decided to try a test installation of the 7.2.0 beta using an external database running on a fast SSD.

For a test system, I set up a VM running digikam-7.2.0-beta1-x86-64.appimage on Ubuntu 20.04.1 LTS, Linux 5.4.0-58-generic.  I configured digikam to use an external MariaDB Server 10.3.25.  The system, databases, and image lib are on an M.2 connected,  SSD.  This is a clean installation, everything is up to date, with no OS additions other than to install MariaDB Server. I copied  ~24, 000 images from my image library to a new directory on the test system.

I am really impressed with the performance improvement.  All functions run dramatically faster.  The only thing that is not working as expected is the similarity search.

I use the similarity search a lot on my production installation so I am familiar with how to use it.  On the test system, I duplicated a couple of images,  "Find duplicates" does not find them.  It doesn't matter if the duplicates are in the same directory or separate directories, similarity returns no hits, however,  I can see activity on the similarity database when the find duplicates process is active.

I have tried:

• updating the fingerprints
• increasing the similarity range
• running Tools->Maintenance (all options selected)
• checking the permissions of the digikam user on all databases

I watched the databases during the maintenance and they all seem to be experiencing activity.  Any advice on where to look next would be greatly appreciated.

[Maik Qualmann]

A short test here withe the Beta2 shows no problem, also MySQL database. But
there are no known problems with Beta1 either. You can try digiKam-7.2.0-Beta2
from here: https://files.kde.org/digikam/

Maik

Am Freitag, 25. Dezember 2020, 21:58:04 CET schrieb Mark Mara:

> I am currently using digikam 7.1.0, (default database(s). It meets or
> exceeds my needs in all areas except performance. I decided to try a test
> installation of the 7.2.0 beta using an external database running on a fast
> SSD.
>
> For a test system, I set up a VM running digikam-7.2.0-beta1-x86-64.appimage
> on Ubuntu 20.04.1 LTS, Linux 5.4.0-58-generic.  I configured digikam to use
> an external MariaDB Server 10.3.25.  The system, databases, and image lib
> are on an M.2 connected,  SSD.  This is a clean installation, everything is
> up to date, with no OS additions other than to install MariaDB Server. I
> copied  ~24, 000 images from my image library to a new directory on the
> test system.
>
> I am really impressed with the performance improvement.  All functions run
> dramatically faster.  The only thing that is not working as expected is the
> similarity search.
>
> I use the similarity search a lot on my production installation so I am
> familiar with how to use it.  On the test system, I duplicated a couple of
> images,  "Find duplicates" does not find them.  It doesn't matter if the
> duplicates are in the same directory or separate directories, similarity
> returns no hits, however,  I can see activity on the similarity database
> when the find duplicates process is active.
>
> I have tried: