[spambayes-bugs] [ spambayes-Bugs-1166146 ] Tokenizer fails on bad
URL
SourceForge.net
noreply at sourceforge.net
Tue Mar 29 07:29:45 CEST 2005
Bugs item #1166146, was opened at 2005-03-19 07:27
Message generated for change (Comment added) made by anadelonbrin
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=498103&aid=1166146&group_id=61702
Category: None
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Leonid (leobru)
Assigned to: Tony Meyer (anadelonbrin)
Summary: Tokenizer fails on bad URL
Initial Comment:
The following line in the body of a message being
scored or trained
http://)
causes spambayes to die. The potentially relevant
spambayesrc settings are
[Tokenizer]
x-fancy_url_recognition=True
x-pick_apart_urls=True
[URLRetriever]
x-slurp_urls=True
The relevant backtrace is
File
"/usr/home/leob/spambayes-2/spambayes/classifier.py",
line 374, in _add_msg
for word in Set(wordstream):
File "/usr/local/lib/python2.3/sets.py", line 429, in
__init__
self._update(iterable)
File "/usr/local/lib/python2.3/sets.py", line 383, in
_update
for element in iterable:
File
"/usr/home/leob/spambayes-2/spambayes/classifier.py",
line 762, in _add_slurped
slurped_tokens = self._generate_slurp()
File
"/usr/home/leob/spambayes-2/spambayes/classifier.py",
line 556, in _generate_slurp
tokens = self.slurp(*slurp_wordstream)
File
"/usr/home/leob/spambayes-2/spambayes/classifier.py",
line 663, in slurp
domain = mo.group(1)
AttributeError: 'NoneType' object has no attribute 'group'
----------------------------------------------------------------------
>Comment By: Tony Meyer (anadelonbrin)
Date: 2005-03-29 17:29
Message:
Logged In: YES
user_id=552329
Ah - turns out that the -o command line option doesn't work
with the slurp options (options change is made too late), so
I wasn't actually turning slurping on before, which is why I
couldn't duplicate this.
I've identified the cause of this (I haven't checked, but
http://: and various others should also trigger it), and
checked in a fix.
Thanks!
----------------------------------------------------------------------
Comment By: Leonid (leobru)
Date: 2005-03-29 16:46
Message:
Logged In: YES
user_id=790676
NB:
http://(
with the OPENING parenthesis works ok, it's the CLOSING one
that causes the error.
----------------------------------------------------------------------
Comment By: Leonid (leobru)
Date: 2005-03-29 16:45
Message:
Logged In: YES
user_id=790676
I've downloaded version 1.0.4 and did the following:
% sb_filter.py -d dummy -n
% sb_filter.py -d dummy
http://)
^D
And got:
Traceback (most recent call last):
File "./sb_filter.py", line 257, in ?
main()
File "./sb_filter.py", line 248, in main
action(msg)
File "./sb_filter.py", line 180, in filter
return self.h.filter(msg)
File
"/usr/home/leob/spambayes-1.0.4/scripts/spambayes/hammie.py",
line 109, in filter
prob, clues = self._scoremsg(msg, True)
File
"/usr/home/leob/spambayes-1.0.4/scripts/spambayes/hammie.py",
line 38, in _scoremsg
return self.bayes.spamprob(tokenize(msg), evidence)
File
"/usr/home/leob/spambayes-1.0.4/scripts/spambayes/classifier.py",
line 246, in slurping_spamprob
slurp_tokens = list(self._generate_slurp())
File
"/usr/home/leob/spambayes-1.0.4/scripts/spambayes/classifier.py",
line 559, in _generate_slurp
tokens = self.slurp(*slurp_wordstream)
File
"/usr/home/leob/spambayes-1.0.4/scripts/spambayes/classifier.py",
line 689, in slurp
domain = mo.group(1)
AttributeError: 'NoneType' object has no attribute 'group'
----------------------------------------------------------------------
Comment By: Tony Meyer (anadelonbrin)
Date: 2005-03-23 17:57
Message:
Logged In: YES
user_id=552329
The problem is with the x-slurp_urls option from the trace,
but I can't duplicate this by adding in "http://)" to a
message. Do you have a message that triggers this problem
that you can attach here for me to test against?
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=498103&aid=1166146&group_id=61702
More information about the Spambayes-bugs
mailing list