tinytim-discuss — Discussions relating to tinyTiM

Re: [tinyTiM-discuss] Fwd: importing/exporting RDF

[Lars H. <he...@se...>]

Hi Lukas,

[...]
> Isidorus (when the mapping concepts are not absolutely different). So I
> started to import data with tinyTiM - but importing RDF/XML reqieres a
> mapping-schema. Are any example schemata existing?

No, the schema is always source-dependent. One could define a standard
mapping which maps the RDF type statement to the tmdm:type-instance
association, but this is not done yet. The RDF readers implement the
RTM mapping described at [1].

> Is it possible to import "normal" RDF/XML files without predifend schemata?

No, since the RDF importers are forced to ignore all statements for
which no mapping is defined, the resulting topic map would always be
empty.

If Isidorus supports CXTM, you can use the CXTM tests provided by the
CXTM test suite [2]. Note, that the last release does not contain the
RDF tests, you have to use the the trunk [3].

All RDF tests are self-contained, which means that the mapping is
defined within the RDF source.


> What's about exporting Topic Maps as RDF/XML - I didn't find a 
> Class/Interface which is able to export TM as RDF/XML?

tinyTiM does not contain any RDF writer yet. I don't have any plans to
write one, since the demand seems to be non-existent. If there are
enough requests, I'll write one. Ontopia provides a RDF/XML writer,
though.


[1] <http://www.ontopia.net/topicmaps/materials/rdf2tm.html>
[2] <http://cxtm-tests.sourceforge.net/>
[3]
<http://cxtm-tests.svn.sourceforge.net/viewvc/cxtm-tests/trunk/rdf/>

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

[tinyTiM-discuss] Fwd: importing/exporting RDF

[Lars H. <he...@se...>]

Moved to the correct mailing list.



This is a forwarded message

From: Lukas Gießmann <in...@Fh...>
To: tin...@li... <tin...@li...>
Date: Sunday, August 23, 2009 1:27:03 PM
Subject: [tinyTiM-commits] importing/exporting RDF

===8<==============Original message text===============

Hi,

I am implementing an RDF importer and exporter for the TM engine 
Isidorus. I also want to get a brief introduction into the RDf-mapping 
in tinyTiM, to be able to exchange mapped data between tinyTiT and 
Isidorus (when the mapping concepts are not absolutely different). So I 
started to import data with tinyTiM - but importing RDF/XML reqieres a 
mapping-schema. Are any example schemata existing?
Is it possible to import "normal" RDF/XML files without predifend schemata?
What's about exporting Topic Maps as RDF/XML - I didn't find a 
Class/Interface which is able to export TM as RDF/XML?


Thanks & Regards

Lukas

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
tinyTiM-commits mailing list
tin...@li...
https://lists.sourceforge.net/lists/listinfo/tinytim-commits



===8<===========End of original message text===========

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

[...]
> Solving <http://code.google.com/p/tmapix/issues/detail?id=7> should
> generate less errors, though.

I created a TMAPIX i/o snapshot (0.3.0):
<http://code.google.com/p/tmapix/downloads/list> which contains the
same set of readers as tinyTiM and the XTM 1.0 / 2.0 and JTM writer.

The XTMTopicMapReader / XTM10TopicMapReader have a method to disable /
enable the translation to TMDM (disabled by default). If the user
disables the translation, the readers work with TMAPI 1.0/tinyTiM 1.5
without exceptions.

The tests enable the translation, though. So you'll get the same set
of failures / errors if you try ``ant test.tinytim-1``. But it works
if you don't enable the translation in your app.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

[...]
> Anyway, TMAPI 1.0 has no priority for me, so it's unlikely that I put
> much effort into fixing the remaining errors. The tinyTiM 1.5 tests
> are not executed by default (only Ontopia, tinyTiM 2.x, generic
> TMAPI), you can test it with

>         ant test-tinytim-1

Now:
    ant test.tinytim-1

:)

Solving <http://code.google.com/p/tmapix/issues/detail?id=7> should
generate less errors, though.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

[...]
>> For those who want to support both TMAPI 1.x and 2.x backends, this
>> shouldn't change much anyway, right? (So, actually, my vote is +1)

> I didn't test it with tinyTiM 1.5 / TMAPI 1, but it may work. You
> can't use the TMAPIX writers, of course. They use the TMAPI 2.0
> interfaces.

Testing revision 85 against tinyTiM 1.5/TMAPI 1.0 caused 714 errors
(running 714 tests) which is a bit, hmmm, suboptimal.

Revision 86 works partially with tinyTiM 1.5: Failures: 62, Errors:
126. The problem is a helper class that uses TMAPI 2.0 to convert XTM
1.0 PSIs to TMDM PSIs.

Anyway, TMAPI 1.0 has no priority for me, so it's unlikely that I put
much effort into fixing the remaining errors. The tinyTiM 1.5 tests
are not executed by default (only Ontopia, tinyTiM 2.x, generic
TMAPI), you can test it with

        ant test-tinytim-1


Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

Hi Markus,

>> I wonder if I should drop tinyTiM MIO in the near future.
>>   
> To make the decision process a bit more challenging: *-2*

:P

> -- That's what I'd love to say :p, but seriously, while I have yet to
> review/compare the individual TMAPIX interfaces, the more generic they
> are, the better. A short example how to migrate between the two
> libraries in question would be nice (and also an occasion to fill the
> wiki). :)

Yes, there will be documentation, of course. The main change will be
the package name, the interfaces / implementations are direct copies
from tinyTiM, so I don't expect that it causes any trouble to migrate.
IIRC you use semagia-mio directly and bypass tinytim-mio anyway,
right? Then you shouldn't recognize any change.

> For those who want to support both TMAPI 1.x and 2.x backends, this
> shouldn't change much anyway, right? (So, actually, my vote is +1)

I didn't test it with tinyTiM 1.5 / TMAPI 1, but it may work. You
can't use the TMAPIX writers, of course. They use the TMAPI 2.0
interfaces.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Markus U. <mar...@gm...>]

Lars Heuer schrieb:
> I wonder if I should drop tinyTiM MIO in the near future.
>   
To make the decision process a bit more challenging: *-2*

-- That's what I'd love to say :p, but seriously, while I have yet to
review/compare the individual TMAPIX interfaces, the more generic they
are, the better. A short example how to migrate between the two
libraries in question would be nice (and also an occasion to fill the
wiki). :)
For those who want to support both TMAPI 1.x and 2.x backends, this
shouldn't change much anyway, right? (So, actually, my vote is +1)

Ad astra, Markus

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Hannes N. <h.n...@go...>]

Hi Lars,

> Well, I hope TMAPIX I/O is useful for other projects as well. At
> least as long as there is no writer/reader interface in TMAPI. Code
> which uses TMAPI is ideally independent of any particular engine, so
> TMAPIX I/O brings the user closer to that goal than tinyTiM MIO since
> tinytim-mio only knows how to setup the TinyTimMapInputHandler but
> knows nothing about other engines or TMAPI.

Well I hope so too. It wasn't criticism, when I said you wouldn't drop
TMAPIX I/O.
On the contrary I like the portability of it :)

best regards
Hannes

-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

Hi Hannes,

> Well I guess you won't drop TMAPIX IO and maintaining two similar
> libraries is not very efficient.

Well, I hope TMAPIX I/O is useful for other projects as well. At
least as long as there is no writer/reader interface in TMAPI. Code
which uses TMAPI is ideally independent of any particular engine, so
TMAPIX I/O brings the user closer to that goal than tinyTiM MIO since
tinytim-mio only knows how to setup the TinyTimMapInputHandler but
knows nothing about other engines or TMAPI.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Hannes N. <h.n...@go...>]

Well, actually this line is completely unnecessary because i added
TMDM.SUBJECT to the locators regarding Occurrences. (Because it is
used as occurrence-type.)

I thought I already removed it.. obviously not :(

regards
Hannes



On Tue, Aug 18, 2009 at 1:46 PM, Lars Heuer<he...@se...> wrote:
> Hi Hannes,
>
>>> I see. Do you implement it (either (2) or (3))?
>> I will implement (3) tomorrow morning.
>
> Thanks.
>
> I changed a few lines, though:
>
>         // remove tmdm:subject, because filter below doesn't work. tmdm:subject
> plays a role, so
>         // _omitTopic returns false - we definitly don't want "tmdm:subject ." in
> our ctm file, do we?
> -        Topic topic = topicMap.createTopicBySubjectIdentifier(TMDM.SUBJECT);
> -        if ( (topic!=null) && (topics.contains(topic)) ) {
> +        Topic topic = topicMap.getTopicBySubjectIdentifier(TMDM.SUBJECT);
> +        if (topic != null) {
>             topics.remove(topic);
>         }
>
>
> I don't see a reason to create tmdm:subject if it does not exist in
> the topic map. If we create tmdm:subject, the topic will never be
> null.
>
> Committed in rev. 354.
>
> Best regards,
> Lars
> --
> Semagia
> <http://www.semagia.com>
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> tinyTiM-discuss mailing list
> tin...@li...
> https://lists.sourceforge.net/lists/listinfo/tinytim-discuss
>



-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

Re: [tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Hannes N. <h.n...@go...>]

Hi,

As long there is an IO library I don't mind how it is called ;)

Well I guess you won't drop TMAPIX IO and maintaining two similar
libraries is not very efficient.

Long story short: I agree. (which means +1 :) )

regards
Hannes

On Tue, Aug 18, 2009 at 1:54 PM, Lars Heuer<he...@se...> wrote:
> Hi all,
>
> I wonder if I should drop tinyTiM MIO in the near future.
>
> I added tinyTiM's JSON Topic Maps writer and the XTM writers to the
> TMAPIX I/O project. That required only 3 - 5 lines code changes since
> the writers do not used much of tinyTiM's internal API. The writers in
> the TMAPIX I/O project operate upon TMAPI and should therefor work
> with any TMAPI 2.0 compatible engine.
>
> The CXTMWriter may require some more work, though; and I did not look
> at the CTMWriter yet.
>
> Anyway, if you look closely at tinyTiM's MIO package you'll notice
> that it contains very few dependencies at tinyTiM (thanks to the
> Streaming API).
>
> I won't remove tinytim.mio tomorrow or next week, I just want to hear
> what the users of tinyTiM think about such a step (yes, I mean you
> both ;)).
>
> Best regards,
> Lars
> --
> Semagia
> <http://www.semagia.com>
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> tinyTiM-discuss mailing list
> tin...@li...
> https://lists.sourceforge.net/lists/listinfo/tinytim-discuss
>



-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

[tinyTiM-discuss] tinyTiM MIO vs. TMAPIX I/O

[Lars H. <he...@se...>]

Hi all,

I wonder if I should drop tinyTiM MIO in the near future.

I added tinyTiM's JSON Topic Maps writer and the XTM writers to the
TMAPIX I/O project. That required only 3 - 5 lines code changes since
the writers do not used much of tinyTiM's internal API. The writers in
the TMAPIX I/O project operate upon TMAPI and should therefor work
with any TMAPI 2.0 compatible engine.

The CXTMWriter may require some more work, though; and I did not look
at the CTMWriter yet.

Anyway, if you look closely at tinyTiM's MIO package you'll notice
that it contains very few dependencies at tinyTiM (thanks to the
Streaming API).

I won't remove tinytim.mio tomorrow or next week, I just want to hear
what the users of tinyTiM think about such a step (yes, I mean you
both ;)).

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Lars H. <he...@se...>]

Hi Hannes,

>> I see. Do you implement it (either (2) or (3))?
> I will implement (3) tomorrow morning.

Thanks.

I changed a few lines, though:

         // remove tmdm:subject, because filter below doesn't work. tmdm:subject
plays a role, so
         // _omitTopic returns false - we definitly don't want "tmdm:subject ." in
our ctm file, do we?
-        Topic topic = topicMap.createTopicBySubjectIdentifier(TMDM.SUBJECT);
-        if ( (topic!=null) && (topics.contains(topic)) ) {
+        Topic topic = topicMap.getTopicBySubjectIdentifier(TMDM.SUBJECT);
+        if (topic != null) {
             topics.remove(topic);
         }


I don't see a reason to create tmdm:subject if it does not exist in
the topic map. If we create tmdm:subject, the topic will never be
null.
         
Committed in rev. 354.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Hannes N. <h.n...@go...>]

> I see. Do you implement it (either (2) or (3))?
I will implement (3) tomorrow morning.

regards
Hannes

-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Lars H. <he...@se...>]

Hi Hannes,

[...]
>> (2) and (3) are okay for me.
>>
>> Do you mean with (3) that this is done in the CTMWriter and not in the
>> preprocessor?

> actually I meant both to be in the CTMWriter, where all the other
> filtering is done.

I see. Do you implement it (either (2) or (3))?

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Hannes N. <h.n...@go...>]

Hi LArs,

> (2) and (3) are okay for me.
>
> Do you mean with (3) that this is done in the CTMWriter and not in the
> preprocessor?

actually I meant both to be in the CTMWriter, where all the other
filtering is done.

regards
Hannes


> Best regards,
> Lars
>
> --
> Semagia
> <http://www.semagia.com/>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> tinyTiM-discuss mailing list
> tin...@li...
> https://lists.sourceforge.net/lists/listinfo/tinytim-discuss
>



-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

Re: [tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Lars H. <he...@se...>]

Hi Hannes,

> testing the CTM-export of Onotoa I realized a bug (or feature) of the
> CTM serializer.

It's a feature, of course. :)

[...]
> 1. Let these topic be in the CTM file, which means having an entry
> like: "tmdm:subject ." - definitly not what I prefer.
> 2. Filter these types ignoring the association in an additional filter method.
> 3. Always ignore associations where topics of the standards are used.

(2) and (3) are okay for me.

Do you mean with (3) that this is done in the CTMWriter and not in the
preprocessor?

Best regards,
Lars

-- 
Semagia
<http://www.semagia.com/>

[tinyTiM-discuss] tmdm:subject/tmdm:name filtering in CTMWriter

[Hannes N. <h.n...@go...>]

Hi everybody,

testing the CTM-export of Onotoa I realized a bug (or feature) of the
CTM serializer.
The topics, using the PSIs of the standard, like tmcl:topic-type, will
be filtered, so they are not
int the CTM file. These topics only will be filtered, if they only
have one subject identifier and
participate in any association. This works for the topic-types,
because the type-instance association
is hidden by the addTypes method in TMAPI.

However there is a problem if tmdm:subject or tmdm:topic-name is used
in a TMCL constraint. Because they are
explicitly in an association. Long story short, the question is, how
do we solve this problem. There are a few options:

1. Let these topic be in the CTM file, which means having an entry
like: "tmdm:subject ." - definitly not what I prefer.
2. Filter these types ignoring the association in an additional filter method.
3. Always ignore associations where topics of the standards are used.

I would prefer 3. Even if the types are used in an association it is
not necessary to specify them before. CTM creates topics on demand, if
I remember correctly.


What do you think?

regards
Hannes

-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

[tinyTiM-discuss] TMAPIX I/O as alternative to tinyTiM MIO

[Lars H. <he...@se...>]

Hi all,

Just a short notice that I created a TMAPIX I/O package which works
similar to tinyTiM's mio subproject (although TMAPIX I/O does not
contain any topic map writer implementation).

That package uses the Streaming Topic Maps API and uses either a
generic TMAPI implementation of the IMapHandler or a specialised
IMapHandler (like tinytim.mio.TinyTimMapInputHandler) if it detects
the underlying TMAPI implementation.

So, if you mainly read topic maps with a TMAPI 2.0 compatible Topic
Maps engine, you may want to give TMAPIX I/O a try since it should
work with any TMAPI 2.0 implementation (like Ontopia).


<http://code.google.com/p/tmapix/downloads/list>

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] tinyTiM distribution

[Lars H. <he...@se...>]

Hi Markus,

[...]
> I like that naming scheme:
>  - org.tinytim.mio.something.ugly    <-- this is where workarounds go
>  - org.tinytim.mio.somethingelse    <-- the place for awesome code
>  - org.tinytim.mio.differentthing    <-- things the authors don't agree
> with/grok fully

Hmm..... I've to think about it... ;)

Maybe I should add another distribution: "ueberall-chaos" ... ;)

Best regards,
Lars
-- 
Semagia
<http://www.semagia.com/>

Re: [tinyTiM-discuss] tinyTiM distribution

[Markus U. <mar...@gm...>]

Lars Heuer schrieb:
> [...] so I move these packages into
> org.tinytim.mio.something
I like that naming scheme:
  - org.tinytim.mio.something.ugly    <-- this is where workarounds go
  - org.tinytim.mio.somethingelse    <-- the place for awesome code
  - org.tinytim.mio.differentthing    <-- things the authors don't agree
with/grok fully

(Sorry, couldn't resist...)
Ad astra, Markus

Re: [tinyTiM-discuss] tinyTiM distribution

[Lars H. <he...@se...>]

Hi all,

[...]
> a) tinytim-mio-batteries-included
> b) meta-tinytim (the engine and mio)

Even if I wrote yesterday, that I'd do (a) first and then (b), I
changed my opinion, that I only do (b). :)

The reason: I want to avoid that tinytim-mio steals the package names
"com.semagia.mio.*" and "org.slf4j.*", so I move these packages into
org.tinytim.mio.something and the com.semagia.mio.IMapHandler is then
called "org.tinytim.mio.something.IMapHandler".

But tinyTiM's IMapHandler implementation belongs to the tinytim.jar
and therefor won't find the repackaged interface IMapHandler.

Are you still following? ;) To cut the long story short: (b) is easier
to achieve than (a), that's all. :)

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] TMCL PSIs

[Lars H. <he...@se...>]

Hi Hannes,

> Good. Already missed it :)

The problem was (and still is), that tmdm:subject is only defined as
glossary term. The newly added TMDM.SUBJECT is currently undefined in
TMDM since it refers to
<http://psi.topicmaps.org/iso13250/model/subject> and *not* to
<http://psi.topicmaps.org/iso13250/glossary/subject>.

The glossary PSIs cause a lot of confusion since people tend to assume
that i.e. <http://psi.topicmaps.org/iso13250/glossary/association> is
the supertype of all TMDM associations, which is not the case.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>

Re: [tinyTiM-discuss] TMCL PSIs

[Hannes N. <h.n...@go...>]

Good. Already missed it :)

regards
Hannes

On Fri, Aug 14, 2009 at 2:24 PM, Lars Heuer<he...@se...> wrote:
> Hi Hannes,
>
>> I will look into it, but not before monday, need to finish some other
>> Onotoa stuff first.
>
> Okay, no problem. BTW, I added tmdm:subject to voc.TMDM, maybe it's
> useful for you, too.
>
> Best regards,
> Lars
> --
> Semagia
> <http://www.semagia.com>
>
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> tinyTiM-discuss mailing list
> tin...@li...
> https://lists.sourceforge.net/lists/listinfo/tinytim-discuss
>



-- 
Onotoa - Simply create your Topic Maps schemas.
http://onotoa.topicmapslab.de

http://www.topicmapslab.de/people/Hannes_Niederhausen

Re: [tinyTiM-discuss] TMCL PSIs

[Lars H. <he...@se...>]

Hi Hannes,

> I will look into it, but not before monday, need to finish some other
> Onotoa stuff first.

Okay, no problem. BTW, I added tmdm:subject to voc.TMDM, maybe it's
useful for you, too.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>