[Rock-dev] csv_logger and rock-toolchain

[Sylvain Joyeux]

On 11/13/2013 09:46 AM, Matthias Goldhoorn wrote:
> The problem here was that a first version (besides the missing
> functionality regarding automatic handling of input_ports was REALLY
> slow. (replay through ruby scripts). So i went for a alternative
> logger implementation, that of course could be used offline too.
I guess that, if slow, the issue was with the re-ordering of samples.

Now, looking at your csv_logger implementation, I see that you actually
are *not* reordering samples. You dump them as they come.

Well, the order of samples *in the log file* is actually the order of
reception by the logger. In other words, if you were to pass through the
log file in sequence, you would get exactly the functionality that you
have implemented in csv_logger. Not straightforward (as the API is
currently designed for a per-stream access), but pretty easy to do. You
can even do that in C++ using the pocolog_cpp library.

The win ? *Everyone* benefits, we can integrate it in rock-export
(making it easy to integrate in GUIs) instead of having to log through a
separate logger to get CSV, and yet another to get HDF5.
>>
>> I agree, that we lose the information about the connections if syskit
>> is not used and I am open for any discussions to solve this issue in
>> a generic way and not just for the CSV logging.
> Then maybe we should start a discussion about logging the connections
> at this point. I don't want to have influence on too much core
> components for this logging functionality. As usual it needs to be
> done, and i tried the most generic way without too much overhead.
The fact that you are trying something is really not the issue. The
issue is that by putting in rock-toolchain, you make your experiments
de-facto part of the official workflow.

-- 
Sylvain Joyeux (Dr.Ing.)
Senior Researcher

Space & Security Robotics
Underwater Robotics

!!! Achtung, neue Telefonnummer!!!

Standort Bremen:
DFKI GmbH
Robotics Innovation Center
Robert-Hooke-Straße 5
28359 Bremen, Germany

Phone: +49 (0)421 178-454136
Fax:   +49 (0)421 218-454150
E-Mail: robotik at dfki.de

Weitere Informationen: http://www.dfki.de/robotik
-----------------------------------------------------------------------
Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
Firmensitz: Trippstadter Straße 122, D-67663 Kaiserslautern
Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
(Vorsitzender) Dr. Walter Olthoff
Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
Amtsgericht Kaiserslautern, HRB 2313
Sitz der Gesellschaft: Kaiserslautern (HRB 2313)
USt-Id.Nr.:    DE 148646973
Steuernummer:  19/673/0060/3
-----------------------------------------------------------------------