Re: [SMW-devel] Semantic Parser Engine

[Yaron K. <ya...@gm...>]

Okay, wow. :) Indeed, there's no way to do this kind of querying (to the
extent that I can even understand it) using regular SMW. But, in lieu of
writing a lot of new code: have you considered accessing the SMW data as
RDF? There are a whole set of tools for doing complex queries on RDF data,
and there are even some tools that let you run SPARQL queries on a wiki's
data from directly within the wiki. See here for an explanation of how to
export SMW data to RDF:

http://semantic-mediawiki.org/wiki/Help:RDF_export

-Yaron


On Thu, Apr 22, 2010 at 8:55 PM, Nathanael Thompson <tha...@gm...>wrote:

> Here's 2:
> Note, these use cases happen in a perfect (less exciting) world.
>
> Seemless parsing experience based on semantic info put in place by someone
> else:
>
> Someone else previously specified that:
> - UCAG are common tags for nucleic acids
> - what amino acids result from what nucleic acids triplets
> - what triplets are used to start and stop the encoding a protein
> - what amino acid sequences make up each protein
> Another person has a long sequence of nucleic acid values and wants to know
> what proteins it would produce.  She specifies that she is inputting a
> nucleic acid sequence and that she wants proteins outputted.  The parser
> uses the combination of tag info and the info about how to get proteins from
> nucleic acid sequences to seemlessly provide the scientist with the proteins
> she desires.
>
>
> Gaining semantic info from a grammar that can be used in non parser task:
>
> A person writes a grammar for encoding family tree info using ebnf. This
> grammar includes the following lines:
>
> Person = [Spouse], Mother, Father, {Child};
> Spouse = "Spouse: ", Name;
>
> He just enters the ebnf into the wiki because someone else already
> specified how to parse ebnf.  From this text the wiki stores the fact that
> with this grammar "Spouse:" specifies that the name of a spouse is coming.
> Also, the parser stores that a person has 0 or 1 spouses, exactly one mother
> and father and any number of children (All valuable ontology info that can
> be used elsewhere).  After inputting the grammar the person thinks to him
> self: "parser info seems a lot like ontology info plus tag info".  Elsewhere
> ... Droid asks the wiki how many mothers a person has and it answers exactly
> 2.
>
> I'm fine making an extension for this.
>
> Nate
>
>
> On Thu, Apr 22, 2010 at 9:57 AM, Yaron Koren <ya...@gm...> wrote:
>
>> Nathanael - could you present a use case or two for the kind of querying
>> you're talking about?
>>
>> -Yaron
>>
>>
>> On Thu, Apr 22, 2010 at 10:04 AM, Jeroen De Dauw <jer...@gm...>wrote:
>>
>>> Hey,
>>>
>>>
>>> > How likely will Wikipedia adopt SMW some time in the future?
>>> This won't be any time soon. There is some effort going on to get a
>>> stripped down version of SMW that does not have query capabilities onto
>>> commons though. Getting SMW on other WMF wiki's will depend a lot on the
>>> outcome of this.
>>>
>>>
>>> > Would it be best for me to develop this in the main SMW source code, a
>>> branch, or an extension?
>>> An extension, as we do not want to add any more code to SMW core.
>>>
>>> Cheers
>>>
>>> --
>>> Jeroen De Dauw
>>> * http://blog.bn2vs.com
>>> * http://wiki.bn2vs.com
>>> Don't panic. Don't be evil. 70 72 6F 67 72 61 6D 6D 69 6E 67 20 34 20 6C
>>> 69 66 65!
>>> --
>>>
>>>
>>> On 22 April 2010 07:19, Nathanael Thompson <tha...@gm...> wrote:
>>>
>>>> Hello All,
>>>>
>>>> I'm interested in adding a feature to Semantic MediaWiki, and I'd like
>>>> to hear what the community thinks about it.  But first, I was wondering: How
>>>> likely will Wikipedia adopt SMW some time in the future?
>>>>
>>>> I'm interested in creating a parser that uses ontologies to help with
>>>> parsing.  For example, a functional property might be used like a regular
>>>> expression "?" and no upper cardinality restriction could be used like a
>>>> regular expression "*".  Also, tag info could be used to help the parser
>>>> know which property to use next while parsing.  Of course, the
>>>> implementation would be much more complicated than I can describe in an
>>>> email.  I think this type of parser would be useful because it will allow
>>>> people who have a bunch of data in text format to easily set up the grammar
>>>> for the data and integrate it into the SMW database.
>>>>
>>>> How well would this mesh with the rest of the work going on in SMW?
>>>> Would it be best for me to develop this in the main SMW source code, a
>>>> branch, or an extension?  Is there anybody else who would like to
>>>> collaborate with this project?  Any other questions or comments would be
>>>> appreciated.
>>>>
>>>> Nate
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> Semediawiki-devel mailing list
>>>> Sem...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/semediawiki-devel
>>>>
>>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Semediawiki-devel mailing list
>>> Sem...@li...
>>> https://lists.sourceforge.net/lists/listinfo/semediawiki-devel
>>>
>>>
>>
>>
>> --
>> WikiWorks · MediaWiki Consulting · http://wikiworks.com
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Semediawiki-devel mailing list
>> Sem...@li...
>> https://lists.sourceforge.net/lists/listinfo/semediawiki-devel
>>
>>
>


-- 
WikiWorks · MediaWiki Consulting · http://wikiworks.com