[Nepomuk] Latency and file list

[Rainer Dorsch]

Artem,

many thanks for your quick and very useful reply.

Am Monday, 8. August 2011 schrieben Sie:
> > Can I get a list of files, which is in my strigi index? I am aware of
> > ~/.kde/share/config/nepomukstrigirc but I would rather want to see a log
> 
> or
> 
> > index dump which confirms what really went in? Also is a way to determine
> 
> how
> 
> > much data from a certain part of the file system went into the index?
> > This would be useful for me to understand, where it is worth to finetune
> > the
> 
> search
> 
> > patterns in ~/.kde/share/config/nepomukstrigirc ...
> 
>  One possible way is cmd: nepomukcmd query "select ?r ?u where { ?r nie:url
> ?u }". It will give you all resources with URL. You need only file:// urls.
> Most likely all of them added by strigi. There exists more exact query if
> you need them.

I did not have nepomukcmd defined on my system, but found on 

http://techbase.kde.org/Development/Tutorials/Metadata/Nepomuk/TipsAndTricks#nepomukcmd

a definition

alias nepomukcmd="sopranocmd --socket `kde4-config --path socket`nepomuk-socket 
--model main --nrl"

which worked for me.

Indeed I got a list of 24k files

rd at blackbox:~$ nepomukcmd query "select ?r ?u where { ?r nie:url?u }"|grep  
file:| wc -l
Total results: 36227
Execution time: 00:00:00.2
24022
rd at blackbox:~$ 

In system settings, Desktop Search displays that more than 30k files are 
indexed. Does anybody know where the difference >30k vs 24k files comes from?

I just found the tutorials on 

http://nepomuk.kde.org/node/2

very nice!

Is 
http://techbase.kde.org/index.php?title=Development/Tutorials/Metadata/Nepomuk/AdvancedQueries 
the best description to start understanding the query langugage?
 
> What does 'how much data' mean ? What is the mesurement unit - RDF
> statement

System Settings tells me, that I have a Nepomuk Store Size of 1.2 GB. That 
seems too much to me... I would be interested to see how these 1.2 GB break 
down. E.g. I know that I have many images indexed, maybe I should take them 
out of the index (the tags probably come from the digikam database(?) not even 
the files themselves).

Many thanks,
Rainer

-- 
Rainer Dorsch
http://bokomoko.de/