[TM4J-users] Re: Problems parsing tolog queries

[Bret C. <bre...@ya...>]

Kal,

A belated "thank you" for your help getting queries to
parse using source locators. I have used your help to
get it working.

FYI, I've also attached my own analysis of how I made
the mistakes I did & how the documentation might be
improved so that people like me (if there are any :>)
don't make the same mistakes. (It's probably more
information than would be helpful, so you might just
skip to part B. Suggestion: Table Describing Different
Types of Locator.)

Bret
--- tm4...@li... wrote:

> Message: 1
> Date: Tue, 24 Aug 2004 09:42:31 +0100
> From: kal <ka...@te...>
> To: Bret Cohen <bre...@ya...>
> CC: tm4j <tm4...@li...>
> Subject: [TM4J-users] Re: resend: Problems parsing
> tolog queries
> 
> Hi Bret,
> 
> If you want to query by topic id, use @id instead of
> the bare id in your 
> query (e.g. @activity).
> 
> However, your assumption about source locators is
> not correct. The 
> source locator for a topic is set by calling 
> topic.addSourceLocator(loc), not
> topic.setSubject(loc). And when the 
> processor resolves a string like 'activity' as a
> source locator it 
> actually resolves it relative to the baseAddress of
> the topic map so you 
> would need to do both of these steps:
> 
> Locator baseLoc =
> tm.getLocatorFactory().createLocator("URI", 
> "file:///sass.xtm");
> tm.setBaseLocator(baseLoc);
>
topic.addSourceLocator(tm.getBaseLocator().resolveRelative("#activity"));
> 
> The first two lines you only need to do once of
> course, and then only if 
> you are not reading the topic map in from a file (if
> you read it in from 
> a file the base address then it should get a
> baseLocator set automatically).
> 
> The third line you should do once and only for
> topics that you create 
> (the ones parsed from the topic map will have a
> source locator set 
> automatically).
> 
> Hope this helps!
> 
> Cheers,
> 
> Kal
> Bret Cohen wrote:
> 
> >This is a resend. The last time I sent this, there
> >were problems that prevented it from going through
> >
> >Kal,
> >
> >Once more, a problem, that once more looks like a
> bug,
> >but almost surely isn't - yet may be a stimulant to
> >improve the documentation. I've attached a file
> >describing my dealings with the problem.
> >
> >It has to do with creating a prepared query when
> the
> >TopicMap has *not* been read in from an XTM file. I
> >have no problem when creating & executing prepared
> >queries on TopicMaps read in from an XTM file.
> >
> >Bret
> >
> >
> >
> >
> >		
> >_______________________________
> >Do you Yahoo!?
> >Win 1 of 4,000 free domain names from Yahoo! Enter
> now.
> >http://promotions.yahoo.com/goldrush
> >
>
>------------------------------------------------------------------------
> >
> >Tolog Query & Source Indicators
> >
> >It appears that Tolog Queries rely on source
> locators
> >(rather than, say Topic IDs)
> >for finding typing topics used within the query.
> >
> >This has caused me some problems, and I still
> haven't
> >figured out the solution. (I thought I did, but
> >it doesn't work.)
> >
> >Two possible remedies:
> >1. Can I or you change the parser so that it can
> >   use Topic IDs instead of source locators to
> locate the typing topic?
> >2. Can you add documentation on creating Tolog
> prepared queries
> >   that deals with the source locator problem?
> >
> >Here's my story - together with the still
> unresolved problem:
> >
> >I had never worked with source locators, until I
> came upon this
> >Exception when I tried to create a prepared query:
> >
> >[java] Exception processing all activities query:
> org.tm4j.tologx.TologParserException: No topic in
> topic map with source locator SAAS.xtm#activity
> >
> >I never had this problem when I was working from a
> loaded
> >XTM file, rather than from a TopicMap created from
> scratch
> >from memory.
> >
> >I then realized that the the readTopicMap() method
> I had been using
> >must have added source locators created by
> appending #topicID
> >to the name of the XTM file from which the data was
> loaded -
> >and that that was why queries run after reading a
> topic map
> >from an XTM file were parsing correctly.
> >
> >So I figured I could create the same effect for a
> TopicMap
> >created from scratch within memory.
> > 
> >So I used the name of the XTM file to which I want
> to later
> >write the TopicMap as my "dummy base locator" for
> the TopicMap
> >& then appended #topicID to create my source
> locators,
> >mimicing the behavior which the readTopicMap()
> method
> >was following.
> >
> >Here's an example:
> >I created a typing topic for a tolog query with ID:
> "activity"
> >I added a src locator as follows (exception
> processing left out)
> >(Actually, I did this using a createTopic() method,
> > but the result should be the same):
> >Topic topic = topicMap.getTopicByID("activity");
> >String baseAddress =
> tm.getBaseLocator().getAddress()+"#"+topic.getID();
> >srcLoc =
> tm.getLocatorFactory().createLocator("URI",
> baseAddress);
> >topic.setSubject(srcLoc);
> >
> >I then printed the source locator for the activity
> topic
> >and the String address obtained from the locato as
> follows:
> >
> >Topic activity = tm.getTopicByID("activity");
> >Locator srcLoc = activity.getSubject();
> >System.out.println("srcLoc of activity topic:
> "+srcLoc);
> >
> >String srcAddr = srcLoc.getAddress();
> >System.out.println("srcAddr of activity topic:
> "+srcAddr);
> >
> >The results indicated that everything went as I had
> planned...
> 
=== message truncated ===



		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail