[spambayes-dev] removing punctuation redux

[Tony Meyer]

Hi Remi,

> If you tell me the command to use I will do a run during my 
> dinner time.
> 
> I'm a bit confused with split.py splitndir, tim timcv etc.....

You should be able to copy Skip's script (at the bottom of his message) and
use that (you're on a *nix system, yes?).

Otherwise, what you basically do is (assumes you have the test directories
all set up like in README-DEVEL):

1.  Run "timcv.py -n 10 -s 12345 > std.txt" (10 is the number of directories
you have)
2.  Make the change you want to test (for Skip's change the punctation
option).
3.  Run "timcv.py -n 10 -s 12345 > other.txt" (10 is the number of
directories you have)
4.  Run "rates.py std.txt"
5.  Run "rates.py other.txt"
6.  Run "cmp.py stds.txt others.txt > results.txt"

The results.txt file will have the results that you're after.

[Note, however, that given the number of times Tim has had to chastise
<wink> me for poor testing methodology, this should be taken with a grain of
salt.]

=Tony Meyer