Hi James, Many thanks for the info and pointers! It gives me a sense of the current thinking in the topic, which is what I was looking for.
On 2013-10-21 18:24, James HK wrote: > Hi, > > ## SMWWriter/POM > Looking at the extension and taken into the account the following > statement "This page describes an extension developed in 2010, which > was not released and appears to have been abandoned." [1] that > SMWWriter/POM will be less likely a candidate for building a core > component and therefore other means [2] should be discussed to support > such feature in future. > > Start and endpoint for SMW is the subject (a wikipage article), any > modification should go through and triggered by an appropriate update > of a subject. If possible all updates should use the official MW API > to avoid any issues that would arise by a data model change (either on > the side of MW or SMW). Yes, this is what is done already in RDFIO. The MW API is available both as a REST service, and as a programmatic API (which is the one that RDFIO uses), which might cause a little confusion in the discussion :) ... so, RDFIO uses MW API for writing (After the modifications of the wiki text is done with the help of WikiObjectModel / WOM). But what would be nice, is to also expose the current SPARQL based fact editing features that RDFIO provide, via the *REST* part of the API, as a separate "API-extension", or what to call it (don't know exactly how this works, but I think that is what SMWWriter did at the time). > ## RDF/ Export > In the past, RAP[3] was part of SMW which has been discontinued for > some reasons and Wikidata [4] decided to use easyRDF for the export. Yeah, EasyRDF seems to be what most people go for. It's functionality does not seem to match ARC2 completely though, which made me wait with a switch to it so far. > ## Serialization > Maybe of interest (while not directly related to this thread), SMW 1.9 > comes with a full serializer/deserializer of a SemanticData container > ([5], [6], [7], [8]) that can be used externally. Interesting! While it does not seem to contain any conversion to/from RDF / SPARQL formats, it might potentially be useful for extensions doing that, such as RDFIO. Will look closer into that. Cheers // Samuel > ## SPARQL > I can't say anything about it as I'm not sufficiently involved in this topic. > > [1] http://semantic-mediawiki.org/wiki/Help:SMWWriter > > [2] http://www.mediawiki.org/wiki/API:Edit > > [3] http://www.wiwiss.fu-berlin.de/suhl/bizer/rdfapi/ > > [4] http://lists.wikimedia.org/pipermail/wikidata-tech/2013-May/000005.html > > [5] > https://github.com/wikimedia/mediawiki-extensions-SemanticMediaWiki/blob/master/includes/api/ApiBrowse.php > > [6] > https://github.com/wikimedia/mediawiki-extensions-SemanticMediaWiki/blob/master/docs/api.md > > [7] > https://github.com/wikimedia/mediawiki-extensions-SemanticMediaWiki/blob/master/includes/serializer/Serializers/SemanticDataSerializer.php > > [8] > https://github.com/wikimedia/mediawiki-extensions-SemanticMediaWiki/blob/master/includes/serializer/Deserializers/SemanticDataDeserializer.php > > Cheers > > On 10/22/13, Samuel Lampa <samuel.la...@gmail.com> wrote: >> Yeah, sorry, you're right. I'm too indulged in my own thoughts, so I >> forget to provide context ... >> >> So, a quick overview of the motivation, history and current status of >> RDFIO: >> >> == Two usa cases == >> >> The start of RDFIO was when me and Denny realized we were thinking of >> the same thing: >> "Providing a means of updating the (semantic) text in wiki articles, >> using SPARQL." >> >> I aditionally wanted to be able to: >> "Import *plain RDF triples*, with no need for an ontology, into wiki >> pages, keep the original URI:s, and be able to export / query the >> triples in their *original* format" >> >> This was not possible at the time, since the existing "ontology import", >> would AFAIK only import the "classes" of an OWN ontology. Not the >> "instances" of plain RDF triples. >> >> >> == Low hanging fruit, using existing parts == >> >> As most pieces was there already to implement these two features, it >> seemed kind of a low hanging fruit. At that point, the existing pieces >> were: >> >> * SMWWriter >> Denny's extension for updating semantic text in wiki articles, via MW API >> >> * PageObjectModel >> An extension providing a programmatic API to access and modify content >> of wiki pages >> >> * ARC2 >> A PHP RDF library, providing parsers between most common RDF formats and >> SPARQL, and also including a pure PHP based SPARQL endpoint. >> >> Later, the SMWWriter and PageObjectModel were dropped, since >> WikiObjectModel could do the same (except expose the functionality via >> MW API), and provided better stability. >> >> == Current status == >> >> In the first version (the one currently in the MW repos), RDFIO was a >> large and ugly hack, and almost impossible to maintain. >> >> Since a year or two back, I've been working on a complete rewrite of >> RDFIO (currently hosted at github) with a much more straightforward >> approach, and with a bit less dependencies (and hope to improve that >> still). >> >> I have now, the last weeks, tried to improve the reliability of the code >> by adding a bunch of unit tests, and some system tests, and I think it >> is now stable enough for wider testing. >> >> === Current features === >> >> The current features are: >> >> 1. RDF Import - Import plain RDF triples into wiki pages and facts, with >> some configurable logic for how to figure out suitable wiki titles for >> URI:s. >> >> 2. SPARQL endpoint - A Pure PHP SPARQL endpoint with write capabilities >> (using ARC2:s custom SPARQL+ syntax) to update wiki article content. >> Also has options to query by using original URI:s (those used when >> importing external RDF data), rather than SMW:s internal format (on eht >> form "http://...Special:URIResolver/...";). >> >> 3. SPARQL Import - Point at an external SPARQL endpoint and start >> importing it's triples (Only manually, 10 triples at a time, so far). >> >> >> >> Hope it provide the context needed! ... otherwise, feel free to ask more! >> >> Best >> // Samuel >> >> >> On 2013-10-21 16:53, Yaron Koren wrote: >>> Hi Samuel, >>> >>> Perhaps a bit of background information would be helpful. What does >>> RDFIO do? What, if anything, do you want it do further? What did >>> SMWWriter do? How do POM/WOM fit into this whole thing? And how does >>> this relate to the RDF triplestore functionality that SMW now has? >>> >>> -Yaron >>> >>> >>> On Mon, Oct 21, 2013 at 10:30 AM, Samuel Lampa <samuel.la...@gmail.com >>> <mailto:samuel.la...@gmail.com>> wrote: >>> >>> Just another note ... future work would probably need to do something >>> about the RDF library dependency situation. The current state of >>> ARC2, >>> with 30 open issues [1], and no SPARQL 1.1 support, is a little >>> worrying. >>> >>> Haven't got my head around EasyRDF [3] enought to say if it covers >>> all >>> the functionality of ARC2 though, so can't tell. >>> >>> [1] https://github.com/semsol/arc2/issues?state=open >>> [2] https://github.com/semsol/arc2/issues/57 >>> [3] http://www.easyrdf.org/ >>> >>> Best >>> // Samuel >>> >>> On 2013-10-21 15:30, Samuel Lampa wrote: >>> > To give some more background for my thoughts, I realize that the >>> main >>> > reasons for not including SMWWriter in core at the time, have >>> been the >>> > dependencies it was relying upon (Page Object Model): >>> > >>> > http://semantic-mediawiki.org/wiki/Help:SMWWriter >>> > >>> > The size of the code is also mentioned, but I wonder whether that >>> is >>> > really a case anymore, with the growing size of the SMW core >>> library >>> > itself anyway? >>> > >>> > I could mention that RDFIO currently depends on Wiki Object >>> Model [1], >>> > rather than Page Object Model, due to better stability, but I had >>> in >>> > my plans to try to drop even that dependency if possible, as the >>> next >>> > step for RDFIO. >>> > >>> > // Samuel >>> > >>> > >>> > On 2013-10-21 15:23, Samuel Lampa wrote: >>> >> Hi all, >>> >> >>> >> What is the current status of (REST) APIs and import >>> functionality in >>> >> SMW? >>> >>> <snip> >>> >> >> ------------------------------------------------------------------------------ >> October Webinars: Code for Performance >> Free Intel webinars can help you accelerate application performance. >> Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most >> from >> the latest Intel processors and coprocessors. See abstracts and register > >> http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk >> _______________________________________________ >> Semediawiki-devel mailing list >> Semediawiki-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/semediawiki-devel >> -- Developer at www.uppmax.uu.se & www.farmbio.uu.se M: samuel.la...@it.uu.se G: samuel.la...@gmail.com S: samuellampa P: +4670-2073732 T: @smllmp B: saml.rilspace.org ------------------------------------------------------------------------------ October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60135031&iu=/4140/ostg.clktrk _______________________________________________ Semediawiki-devel mailing list Semediawiki-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/semediawiki-devel
