-------- Original Message --------
Subject: KIM support
Date: Fri, 26 Jun 2009 14:35:34 +0100
From: <alistair.d...@bt.com>
To: <marin.dimit...@ontotext.com>
Hi Marin,
Hope you're well - long time no see!
John mentioned you might be able to help me get a bit of KIM support.
I've fired a few issues off to the mailing list but generally don't hear
anything back - you must be working those guys too hard!
I've attached them below. Any chance you could bump me up their to-do list?
Thanks,
Alistair
*Alistair Duke
Next Generation Web Research*
*Centre for Information & Security Systems Research *
*BT Innovate
___________________________
Office: +44(0) 1473 608192
Mobile: +44(0) 7730 426257
Fax : +44(0) 1473 606755
Email: alistair.d...@bt.com
--- Begin Message ---
Hi Philip,
Sorry to bombard you with all these KIM issues at the same time, but can
you help the following?
I'm extracting people's names from a list of names where the name is
typically followed by a role:
E.g. Philip Alexiev KIM expert, Alistair Duke KIM novice
Where the people's name are already in my knowledge base then they are
extracted OK but where they are not, the names are extracted as the name
plus the role ie. 'Alistair Duke KIM novice' is a person. I guess this
is to be expected as GATE has no way of knowing how many words are in a
name. So the question then is, how do you tell GATE / KIM to default to
2 words names? I think all of the names I'm extacting are 2 word names.
3 word names tend to be use a hyphen. I guess Dutch and Chinese names
wouldn't fall into this but these are pretty rare / non existant in the
texts.
A related issue is where I have the text:
Alistair Duke & Philip Alexiev
The whole of this is extracted as an organisation but I would prefer 2
people to be extracted. Is there anyway I can control this?
Many thanks,
Alistair
Alistair Duke
Next Generation Web Research
Centre for Information & Security Systems Research
BT Innovate
___________________________
Office: +44(0) 1473 608192
Mobile: +44(0) 7730 426257
Fax : +44(0) 1473 606755
Email: alistair.d...@bt.com
This email contains BT information, which may be privileged or
confidential.
It's meant only for the individual(s) or entity named above. If you're
not the intended
recipient, note that disclosing, copying, distributing or using this
information
is prohibited. If you've received this email in error, please let me
know immediately
on the email address above. Thank you.
We monitor our email system, and may record your emails.
British Telecommunications plc
Registered office: 81 Newgate Street London EC1A 7AJ
Registered in England no: 1800000
--- End Message ---
--- Begin Message ---
Hi,
I have an issue where KIM is finding overlapping entities. In my
extended knowledge base, 'Fred Hall' is a trusted entity of type
SalesTeamIndividual which is a subclass of protont:Person.
However, this is recognised as a facility by KIM - a hall rather than as
a salesman. During annotation, a warning is logged (see below) so I know
at least that KIM knows about Fred the salesman.
I've added a rule in entity-priority.conf:
priority.rule5.name = Preference of SalesTeamIndividual than Facility
priority.rule5.class1 = SalesTeamIndividual
priority.rule5.class2 = Facility
priority.rule5.delta = 300
but this doesn't seem to have any effect (I notice it says that the
priorities are for KIMO classes only).
Is this the correct way to set the priority or do I need to change
something else?
Thanks,
Alistair
Log:
WARN 201612[RMI TCP
Connection(6)-132.146.79.38](LookupOverlapIdentityResolver.java:111) -
Lookup AnnotationImpl: id=5; type=Lookup; features={originalName=James
Hall, class=http://www.bt.com/2008/12/winloss#SalesTeamIndividual,
inst=http://www.ontotext.com/kim/2006/05/wkb#SalesTeamIndividual_T.80323
3865}; start=NodeImpl: id=4; offset=17; end=NodeImpl: id=6; offset=27
exactly overlaps NERC annotation AnnotationImpl: id=829; type=Location;
features={originalName=James Hall, matches=[835, 829],
rule2=FacilityFinal, rule1=Facility1,
class=http://proton.semanticweb.org/2006/05/protonu#Facility};
start=NodeImpl: id=4; offset=17; end=NodeImpl: id=6; offset=27
, but classes don't match.
Alistair Duke
Next Generation Web Research
Centre for Information & Security Systems Research
BT Innovate
___________________________
Office: +44(0) 1473 608192
Mobile: +44(0) 7730 426257
Fax : +44(0) 1473 606755
Email: alistair.d...@bt.com
This email contains BT information, which may be privileged or
confidential.
It's meant only for the individual(s) or entity named above. If you're
not the intended
recipient, note that disclosing, copying, distributing or using this
information
is prohibited. If you've received this email in error, please let me
know immediately
on the email address above. Thank you.
We monitor our email system, and may record your emails.
British Telecommunications plc
Registered office: 81 Newgate Street London EC1A 7AJ
Registered in England no: 1800000
--- End Message ---
--- Begin Message ---
Hi Philip,
Thanks for the reply - I know you're busy.
I've done as you've suggested but unfortunately since it's using the
string rather then the alias id, it also ignores my product line alias
which is also called ICT. I guess the solution is to remove the company
instance called ICT from the wkb (or at least all of it's aliases) and
not bother with the ignore list.
Regards,
Alistair
> -----Original Message-----
> From: Philip Alexiev [mailto:philip.alex...@sirma.bg]
> Sent: 18 June 2009 13:37
> To: Duke,AK,Alistair,CXR3 R
> Subject: Re: [Kim-discussion] Issue with ignore list
>
> Hi Alistair,
>
> Sorry for the late response, we have been very busy lately.
>
> Try listing the aliases themselves in plain text in
> $KIM_HOME/context/default/resources/gazetteer/ignoreList.def
> like this:
>
> ICT Group, Inc.
> ICT Group, Incorporated
> ICT Group
> ICT
>
> Although, beware that this will prevent the gazetteer from
> recognizing the phrase as a known company. There are still
> Jape grammars in KIM which will recognize it as a company.
>
> Greetings,
> Philip
>
>
>
> On 06/09/2009 06:59 PM, alistair.d...@bt.com wrote:
> > Hi Philip,
> >
> > I'm still having problems with the ignore list. I changed
> the entries
> > to include all the aliases of that company (see below) but
> KIM didn't
> > take this into account. One strange thing is that when KIM loads it
> > reports that it's using the ignorelist.def file but gets
> the number of
> > lines wrong i.e. it reported that 11 aliases were found even though
> > I'm only specifying 4:
> >
> > 11 aliases loaded from ignore list at
> > C:\KIM\bin\../context/default/resources/gazetteer/ignoreList.def
> >
> > Can you help?
> >
> > Thanks,
> >
> > Alistair
> >
> >
> <http://www.ontotext.com/kim/2006/05/wkb#PublicCompany_T.270.0><http:/
> > /w ww.ontotext.com/kim/2006/05/kimlo#ignoredAlias> "" .
> >
> <http://www.ontotext.com/kim/2006/05/wkb#PublicCompany_T.270.1><http:/
> > /w ww.ontotext.com/kim/2006/05/kimlo#ignoredAlias> "" .
> >
> <http://www.ontotext.com/kim/2006/05/wkb#PublicCompany_T.270.2><http:/
> > /w ww.ontotext.com/kim/2006/05/kimlo#ignoredAlias> "" .
> >
> <http://www.ontotext.com/kim/2006/05/wkb#PublicCompany_T.270.3><http:/
> > /w ww.ontotext.com/kim/2006/05/kimlo#ignoredAlias> "" .
> >
>
>
--- End Message ---
--- Begin Message ---
Hi again,
I'm trying to use the CoreDB feature of KIM but I'm having a problem
connecting to the oracle database. I've created a new user and
tablespace and pointed KIM at it using the following settings:
com.ontotext.kim.KIMConstants.DOCUMENT_REPOSITORY_TYPE = coredb
com.ontotext.kim.KIMConstants.COREDB_CONNECTION_STRING=jdbc:oracle:thin:
@//localhost:1521/ORCL
com.ontotext.kim.KIMConstants.COREDB_USER=KIM
com.ontotext.kim.KIMConstants.COREDB_PASS=password
However, when I run KIM I get the following error:
Attempt to get connection failed due to: ORA-01017: invalid
username/password; logon denied
And after a few attempts the user account gets locked. The fact that I
can then unlock the KIM account and see the error revert to ORA-01017
seems to indicate that the connection string and user details are
correct.
I've tried connecting via my own java class using the same jar file that
KIM uses (ojdbc14-10.2.0.4.0.jar) and it connects with no problem.
I'm running Oracle Enterprise Edition Release 11.1.0.6.0 - 64bit.
Hoping you can help.
Regards,
Alistair
Alistair Duke
Next Generation Web Research
Centre for Information & Security Systems Research
BT Innovate
___________________________
Office: +44(0) 1473 608192
Mobile: +44(0) 7730 426257
Fax : +44(0) 1473 606755
Email: alistair.d...@bt.com
This email contains BT information, which may be privileged or
confidential.
It's meant only for the individual(s) or entity named above. If you're
not the intended
recipient, note that disclosing, copying, distributing or using this
information
is prohibited. If you've received this email in error, please let me
know immediately
on the email address above. Thank you.
We monitor our email system, and may record your emails.
British Telecommunications plc
Registered office: 81 Newgate Street London EC1A 7AJ
Registered in England no: 1800000
--- End Message ---
_______________________________________________
Kim-discussion mailing list
Kim-discussion@ontotext.com
http://ontotext.com/mailman/listinfo/kim-discussion
