Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Sep 20 21:08
    dependabot[bot] labeled #12
  • Sep 20 21:08
    dependabot[bot] labeled #12
  • Sep 20 21:08
    dependabot[bot] opened #12
  • Sep 20 21:04

    dependabot[bot] on maven

    Bump jena-core from 2.11.1 to 4… (compare)

  • Jun 23 20:49
    dependabot[bot] labeled #11
  • Jun 23 20:49
    dependabot[bot] labeled #11
  • Jun 23 20:49
    dependabot[bot] opened #11
  • Jun 23 20:49

    dependabot[bot] on maven

    Bump jetty-server from 9.3.27.v… (compare)

  • Jun 10 15:47
    dependabot[bot] labeled #10
  • Jun 10 15:47
    dependabot[bot] labeled #10
  • Jun 10 15:47
    dependabot[bot] opened #10
  • Jun 10 15:47

    dependabot[bot] on maven

    Bump jetty-servlets from 9.3.27… (compare)

  • May 18 09:51

    dependabot[bot] on maven

    (compare)

  • May 18 09:51

    mielvds on cleanup

    Bump commons-fileupload from 1.… Merge pull request #8 from rawb… (compare)

  • May 18 09:51
    mielvds closed #8
  • May 18 09:51

    dependabot[bot] on maven

    (compare)

  • May 18 09:51

    mielvds on cleanup

    Bump jetty-server from 8.1.11.v… Merge pull request #9 from rawb… (compare)

  • May 18 09:51
    mielvds closed #9
  • May 18 09:48
    dependabot[bot] labeled #9
  • May 18 09:48
    dependabot[bot] labeled #9
tthoeye
@tthoeye
Question: Would it be possible to use TDT as a LOD "redirect" service, e.g. if we define LOD ID's somewhere (.../id/news/123) can we forward the request to TDT and have TDT do content negotiation? Trouble is that HTML representations and LOD representations may have quite different URLs so we should keep quite a big referral table, and preferably also a smart (à la mod_rewrite) redirection for some types
Pieter Colpaert
@pietercolpaert
Hi Thimo, 2 answers for that: the first thing, probably not preferable is creating 303s manually
This is something not yet implemented over here: tdt/core#120
Another answer, a better one I believe, is to hook up e.g., a virtuoso to tdt/triples
then, tdt/triples will dereference any URI that uses the place where tdt/core+tdt/triples is installed as a baseURI
(e.g., for the current situation, if you would have a triple with a subject http://data.gent.be/trees/1 in your virtuoso, and you configure it that way, it will dereference that URI and show that triple)
tthoeye
@tthoeye
Thanks Pieter
Pieter Colpaert
@pietercolpaert
you can read more about that in the paper "painless URI dereferencing with The DataTank: http://2014.eswc-conferences.org/sites/default/files/eswc2014pd_submission_37.pdf
Bonus: you will have a triple pattern fragments interface without further problems ;)
tthoeye
@tthoeye
:)
Pieter Colpaert
@pietercolpaert
tdt/triples source: https://github.com/tdt/triples (any tdt/core installation can directly be extended with tdt/triples) and a demo: http://triples.demo.thedatatank.com/triples/all
tthoeye
@tthoeye
I will look into it, but I think the difficulty may be that we need to be able to 303 to other resources (HTML on the website, non-linked XML/JSON from even different places/domains, etc...)
Pieter Colpaert
@pietercolpaert
We could extend tdt/triples with that quite easily... but the question is whether you want that: only machines will see these links and when machines arrive at non-linked resources by using a HTTP link which is expected to show more useful information in a machine interpretable format their logic might break
what you could do however is just add a link http://id.stad.gent/trees/1 rdfs:seeAlso http://stad.gent/htmlpagesoftrees/1 .
that way, machines will understand that a rdfs:seeAlso means that what's further is just more information in possibly human only readable formats
in HTTP you're almost obliging a redirect, with a rdfs:seeAlso you're indicating that in the end-userinterface it might be interesting to show this link with the identifier
Niels V
@nvdk
dereferencing typically uses content negotiation to differentiate between humans (requesting html) and machines (rdf, json, etc) does the tdt/triples handle this?
Pieter Colpaert
@pietercolpaert
yes, as you can see at e.g., http://triples.demo.thedatatank.com/NS
you can perform content negotiation on that URI :)
Or we also created different URLs for the fdifferent formats: http://triples.demo.thedatatank.com/NS.ttl
Niels V
@nvdk
great :)
but as thimo mentioned, in this case we would need to redirect humans to other data services (eg drupal or other cms)
Pieter Colpaert
@pietercolpaert
I don't find that a good idea as code will break: the "semantics" behind a 303 redirect is that machines will be able to read the data behind it, which will not be the case if I understand correctly
but if you really want it, you can add it as a feature request to tdt/triples
Niels V
@nvdk
specific (curated) html content is often linked to structured rdf data in my experience, the curated content is often of no real use to machines
Pieter Colpaert
@pietercolpaert
often ;-)
Niels V
@nvdk
so no it does not break the semantics, rdf data will be available
Pieter Colpaert
@pietercolpaert
ok, let's take your word for it ;) - add it as a feature request to tdt/triples and we'll help you further on there :)
tthoeye
@tthoeye
Thanks guys
tthoeye
@tthoeye
Ok guys, looks like tdt/triples can work as a solution for all our data related requirements, but still, we are going to need another component that can lookup the human-readable version of a dataset
Could we add this as an installed resource? (please note: I have no idea what I'm talking about) e.g. when the request has an Accept text/html header, can we have TDT do a lookup somewhere and redirect to the relevant page on the website?
Pieter Colpaert
@pietercolpaert
As mentioned over a very serendipitous encounter/lunch: it seems best that for your use cases you use some custom scripts which actually make the link between the Linked Data identifiers and the HTML pages
Nevertheless I would still argue to implement manually added links through the RDF store you'll be using as well. It would be weird to set up yet another database which contains yet another set of links
Niels V
@nvdk
though not necessarily manually we will definitely have links to the html representation in the rdf as well
tthoeye
@tthoeye
@pietercolpaert @nvdk agree with both of you. Thanks for talking this through.
Maybe we should explore how we could build a LOD service component which does smart redirection based on configuration in the RDF store in another track ;)
Pieter Colpaert
@pietercolpaert
Screenshot from 2015-05-04 17:34:59.png
thanks to a small extra functionality in the tdt/core sparql templating system, you can now also support optional parameters with SPARQL templates
which is really nice! A private SPARQL endpoint + The DataTank is becoming a really nice tool to use! I use it in a lot of projects :)
Jan Vansteenlandt
@coreation
\o/
thanks @coreation
hey, no problem ;)
Pieter Colpaert
@pietercolpaert
@coreation To thank you I've opened a couple of new issues! ;)
Jan Vansteenlandt
@coreation
<o/
Sander Meert
@Dobbelsteen
Hey everyone, Sander the intern at Open Knowledge Belgium here
As Pieter mentioned a few weeks ago, I'll be doing a small overhaul of TheDataTank website, but I'm looking for some input from others.
I've organised some ideas on a Question/Answer board (~StackOverflow clone) on which anyone can comment, upvote or ask new questions.
If any of you has some spare time to check it out and give some feedback, that would be awesome!
Link: http://san.rocks/qa/index.php
Pieter Colpaert
@pietercolpaert
Hey guys, can you please help Sander out here? He's going to update TheDataTank.com to whatever we want it to say. Giving your input should only take 5 minutes (promised): http://san.rocks/qa/index.php
Sander Meert
@Dobbelsteen
Ok, since our new QA platform doesn't have that much success, we've decided to keep it simple and bother you guys for input here :)
The first thing I want to ask is, what are the best/most important feature(s) according to you that TDT offers?
Some of these could be the API, front-end and back-end UI, DCAT generation, etc. Shoot anything that comes to mind
KevinCoworks
@KevinCoworks

Hi all! Wasn't sure were to post this but I'm trying it here :)

I want to install the input package for tdt but when adding "tdt/input" : "dev-master" to my composer.json file I'm getting following problem:
"tdt/streamingrdfmapper dev-master requires coreation/arc2 dev-master -> no matching package found."
which seems to come from dev-master having a different arc2 requirement than all the other version (who use semsol/arc2).

I'm not that versed in composer (yet) so if anyone could point me out what to do in such case it would be much appreciated :)

Pieter Colpaert
@pietercolpaert
Hi @KevinCoworks, seems like that repo wasn't published publicly on packagist (which made it undiscoverable). I've published it now though: https://packagist.org/packages/coreation/arc2
So, if you'll retry, it should work. If not, give us a nudge :)