What about making it a hybrid solution? Mongo for entity/page details, a graph DB for "higher-level" metadata/properties to support queries?
Mongo is great for scalability and as a KV store(maybe you can even store docs as JSON-LD), but doesn't have the querying power of a graph DB or RDBMS. Graph DBs are the reverse. Maybe the Graph DB will just have a pointer to the document in Mongo? New geospatial features in the latest releases of these stores should also be considered. PostGIS extension of Postgres is amazing, but MySQL/MariaDB is catching up and even MongoDB now has some limited geospatial capabilities. Geospatial features (proximity, adjacent, polygon selection, etc) can even be used for non-traditional, non-geo concepts that can projected to a 2d space. Just my 2 cents, Joel Sent from my iPhone > On Oct 16, 2013, at 8:28 AM, Toni Hermoso Pulido <toni...@softcatala.cat> > wrote: > > Hi, > > I don't know how feasible could be to make it more engine agnostic. > It's a different programming language involved but, for instance, > pumpio (https://github.com/e14n/pump.io) uses something according to > that philosophy https://github.com/evanp/databank > > 2013/10/16 Yury Katkov <katkov.ju...@gmail.com>: >> I suppose that Mongo have been chosen because of its scalability, right? >> ----- >> Yury Katkov, WikiVote >> >> >> >> On Wed, Oct 16, 2013 at 5:20 AM, david mason >> <vid_semediawiki-de...@zooid.org> wrote: >>> >>> May I suggest that ElasticSearch is considered instead of MongoDB. >>> >>> ElasticSearch is the index engine of the new MediaWiki Search, so end users >>> won't need to set up and support multiple data stores. Like MongoDB it is a >>> document store that natively uses JSON, and is really easy to set up and run >>> (a .deb is available). It's super easy to work with and since it's based on >>> Lucene incredibly powerful for many operations. I've used both Mongo and ES >>> and definitely prefer the latter. >>> >>> They each have their strengths, MongoDB is more of a key value store, ES is >>> more of a search server (though I'd assert it could do the KV stuff adding >>> very useful search operations and no additional infrastructure if using MW >>> search), in either case this seems like it would be a big win in terms of >>> better structured, more accessible data! >>> >>> David >>> >>> >>>> On 15 October 2013 20:08, Jeroen De Dauw <jeroended...@gmail.com> wrote: >>>> >>>> Hey, >>>> >>>> The last release introduced SQLStore3, a partial rewrite of SQLStore2, >>>> improving on the performance of its predecessor. That is not the end of the >>>> story for the SMW query stores though. >>>> >>>> This email was prompted by work MWJames is doing in supporting MongoDB \o/ >>>> https://gerrit.wikimedia.org/r/#/c/88534/ >>>> >>>> For a while now, there have been two items on our Roadmap about utilizing >>>> new libraries I created for the Wikidata project, that are both based on, >>>> and usable by, SMW components. >>>> >>>> * >>>> https://semantic-mediawiki.org/wiki/Roadmap#Make_use_of_DataValues_library >>>> * https://semantic-mediawiki.org/wiki/Roadmap#Make_use_of_Ask_library >>>> >>>> There now is a third such component, which might enable us to get a nice >>>> improvement to our SQLStore without all to much effort. I described this >>>> here: >>>> >>>> https://semantic-mediawiki.org/wiki/Wikibase_QueryEngine >>>> >>>> This is still quite far off, assuming no one else jumps on it, given that >>>> it requires the earlier two items to be finished first. Feedback on the >>>> idea >>>> is however welcome. And awareness of these preliminary plans, or rather >>>> possibilities (I'm not committed to doing this at this point), is good for >>>> those doing or planning to do something related to the SMW storage >>>> infrastructure. >>>> >>>> Cheers >>>> >>>> -- >>>> Jeroen De Dauw >>>> http://www.bn2vs.com >>>> Don't panic. Don't be evil. ~=[,,_,,]:3 >>>> -- >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> 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 >>> >>> >>> ------------------------------------------------------------------------------ >>> 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 >> >> ------------------------------------------------------------------------------ >> 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 > > > > -- > Toni Hermoso Pulido > http://www.cau.cat > > ------------------------------------------------------------------------------ > 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 ------------------------------------------------------------------------------ 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
