[Clam-devel] Re: [PATCH] evaluation framework for different segmentation algorithms

[Roman Goj]

Umm, scratch this last patch...

We decided with David to do the program in python, not C++, as using
python enables us to forget about the schema...

...the only problem was I had absolutely no knowledge of Python :-O (ok,
I fiddled a little with the SConstruct files, but that doesn't count,
right?)

...but was ever any programmer stopped from coding something only
because they didn't know the language?!? Of course not, so - huge thanks
to David for teaching me Python!!!!!!! :-)))

And what better way is there to celebrate learning a new programming
language, then to wait for the birds to wake up and tell 'em about your
happiness? Well, I know no better way ;-)

Well, ok, maybe completing the task at hand is a good way of celebrating
too... so - the newly committed script is *almost* there, at least it
calculates the two main similarity measures - precision and recall,
comparing two files (hardcoded for now)...

but, Rome wasn't built in a day (not even a night in this case), so:
TODO: (more for me than for the list perhaps)
* names of files should be given as parameters, not hardcoded (in the
morning)
* the length of the file should be taken from the pool (in the morning)
** this means some avoiding of calls after the end of arrays I think...
(in the morning)
* write some tests! (I think I'll skip this for now, though... I've been
checking by hand on each step, everything seems to calculate what it
should...)
* some general cleanup (I think I'll skip this for now - post GSoC (so -
in 5 days ;) ) - as everything seems to be working and there are still
some other things to finish...)
* get some sleep... (I think I'll skip this for... nope, wait, I won't
ZZzzzzz.....

cheers!
roman