Feedback and improvements

[Joachim Breitner]

Hi,

Am Freitag, den 19.07.2013, 00:35 -0400 schrieb
kleinempfaenger at hotmail.com:
> First a little error warning, that I received after some days of 
> recording (>700kB)
> 
> familia at familia-Satellite-L845:~$ arbtt-stats
> Processing data 
> [=================================>.......................]  58%
> arbtt-stats: Unsupported TimeLogEntry version tag 0

sounds like the file is corrupted. The most likely cause is that you
have been going forth and back between versions in a strange way, or
that you have hit a really unlucky moment of killing arbtt-capture.

You can run 
$ arbtt-recover
and then see if
$ arbtt-stats --logfile = ~/.arbtt/capture.log.recovered
works. If it does, move that file over to
~/.arbtt/capture.log.recovered.

(see also
http://darcs.nomeata.de/arbtt/doc/users_guide/arbtt-recover.html)

> The problem with the -i parameter:
> 
> familia at familia-Satellite-L845:~$ arbtt-stats -f '$sampleage <24:00'
> Total time per tag
> ==================
> ___________________Tag_|___Time_|_Percentage_
> Program:gnome-terminal |  3m00s |     100.00
>                  active |  3m00s |     100.00
> 
> 
> familia at familia-Satellite-L845:~$ arbtt-stats -i -f '$sampleage <24:00'
> General Information
> ===================
>                         FirstRecord | 2013-07-19 01:55:52.456123 UTC
>                          LastRecord | 2013-07-19 01:57:52.531737 UTC
>                   Number of records |                              3
>                 Total time recorded |                          3m00s
>                 Total time selected |                          3m00s
>     Fraction of total time recorded |                           150%
>     Fraction of total time selected |                           150%
> Fraction of recorded time selected |                           100%
> 
> familia at familia-Satellite-L845:~$ arbtt-stats -f '$sampleage <24:00' -i
> General Information
> ===================
>                         FirstRecord | 2013-07-19 01:55:52.456123 UTC
>                          LastRecord | 2013-07-19 01:58:52.597183 UTC
>                   Number of records |                              4
>                 Total time recorded |                          4m00s
>                 Total time selected |                          4m00s
>     Fraction of total time recorded |                           133%
>     Fraction of total time selected |                           133%
> Fraction of recorded time selected |                           100%
> 
> First case shows, that -f -parameter works. When -i enters (before or 
> after), -f wont be processed.

It does here:

$ arbtt-stats -i
General Information                                                                   
===================
                       FirstRecord | 2009-08-21 10:16:44.951741 UTC
                        LastRecord | 2013-07-19 07:20:17.463614 UTC
                 Number of records |                         616191
               Total time recorded |                  427d21h51m00s
               Total time selected |                  294d08h14m00s
   Fraction of total time recorded |                            30%
   Fraction of total time selected |                            21%
Fraction of recorded time selected |                            69%
$ arbtt-stats -i -f '$sampleage < 24:00'
General Information                                                                   
===================
                       FirstRecord | 2009-08-21 10:16:44.951741 UTC
                        LastRecord | 2013-07-19 07:21:17.526257 UTC
                 Number of records |                         616192
               Total time recorded |                  427d21h52m00s
               Total time selected |                       8h07m00s
   Fraction of total time recorded |                            30%
   Fraction of total time selected |                             0%
Fraction of recorded time selected |                             0%

Note that the filter only affects the “selected” time; the recorded time
will always be everything. In your case it seems that there are only
samples from within the last 24h – maybe due to the corruption above.

> Instead of graphical reports (probably intended as a step towards 
> endusers?) I would propose easier-to-use filters. In conclusion I always 
> come to think about a little gui with quickfilter for arbtt-stats, to 
> get an overview of the recorded data. Maybe graphical reports could be 
> for now made with existing statistics programs, or just 
> Calc-Spreadsheets. I weren't able run the export-function (tried several 
> times), maybe somebody could explain this. So everything still runs over 
> the terminal, but I could experiment about it. The point is to group and 
> scale the data, to generate usefull graphics for different needs.

You can get all reports also as comma separated values, which can be
imorted into spreadsheets, e.g.
$ arbtt-stats --category=Program --output-format=csv

I’m not sure how well the time format works in these programs, though.

> Another improvement would be the automatic(?) separation of the 
> capture.log file (Maybe this is meant by "Possibly more data sources" in 
> Readme?). The program could save and reset capture.log periodically. 
> This could also be done by the export-function (export capture.log every 
> week, month etc.). In this case the whole record would stay intact.

Why would you want to separate the capture.log file? To make it easier
to select a certain timespan? That should be possible with filters.


Thanks for your suggestions,
Joachim
-- 
Joachim Breitner
  e-Mail: mail at joachim-breitner.de
  Homepage: http://www.joachim-breitner.de
  ICQ#: 74513189
  Jabber-ID: nomeata at joachim-breitner.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <https://lists.nomeata.de/pipermail/arbtt/attachments/20130719/b3020d9e/attachment.asc>