massive false negatives

Dan Stromberg strombrg at dcs.nac.uci.edu
Mon May 5 21:02:40 CEST 2003


On Mon, 2003-05-05 at 10:25, David Relson wrote:
> At 01:16 PM 5/5/03, Dan Stromberg wrote:
> 
> >Wow, thanks for the quick response.
> >
> >Do I need to upgrade to 0.12 before running bogolex.sh?
> >
> >I'm getting:
> >
> >tesuji-strombrg> sh bogolex.sh < bad > bad.mc
> >Option -w requires an argument.
> >bogolex.sh: line 12: 11307 Done                    $BOGOLEXER -p $*
> >      11308 Broken pipe             | sort -u
> >
> >tesuji-strombrg> cat bad.mc
> >
> >tesuji-strombrg>
> 
> Dan,
> 
> No, you don't need 0.12 to generate a .mc count file.  The script uses 
> bogolex to get tokens from the message, then bogoutil to get the counts, 
> then awk to write the formatted .mc file.  To _score_ a message in .mc 
> format _does_ need 0.12 (since there're lexer changes involved).
> 
> bogolex.sh needs $BOGOFILTER_DIR to be set.  I bet you haven't set it.


I didn't set those vars, but the commands are on my $PATH.

tesuji-strombrg> sh -x bogolex.sh < /tmp/bad > /tmp/bad.mc
+ '[' -z '' ']'
++ which bogolexer
+ BOGOLEXER=/dcs/packages/bogofilter/bin/bogolexer
+ '[' -z '' ']'
++ which bogoutil
+ BOGOUTIL=/dcs/packages/bogofilter/bin/bogoutil
+ '[' -z /dcs/packages/bogofilter/bin/bogolexer -o -z
/dcs/packages/bogofilter/bin/bogoutil ']'
+ echo .MSG_COUNT
+ /dcs/packages/bogofilter/bin/bogoutil -w
+ awk 'NF == 3 { printf( "\"%s\" %s %s\n", $1, $2, $3 ) } '
+ /dcs/packages/bogofilter/bin/bogolexer -p
+ sort -u
Option -w requires an argument.
bogolex.sh: line 12: 11822 Done                    $BOGOLEXER -p $*
     11823 Broken pipe             | sort -u

> David
> 
> P.S.  From your machine name, I'd guess that you speak Japanese and/or play 
> Go...

I play go.  I only wish I could read and speak Japanese, in no small
part so I'd have access to Japanese and Chinese go books.  I just got
back from a two-day go tournament in LA last night.  Do you play?  KGS I
hope.  :)

-- 
Dan Stromberg DCS/NACS/UCI <strombrg at dcs.nac.uci.edu>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://www.bogofilter.org/pipermail/bogofilter/attachments/20030505/d08ff846/attachment.sig>


More information about the Bogofilter mailing list