Mmmm - thinking that this might be quite a good match for http://friendfeed.com/cameron... - Cameron Neylon

Yes, was thinking the same thing ... egonw:someThought owl:sameAs cameron:someThought - Egon Willighagen

So I haven't written the messaging bit up yet but the idea is that there could be a generic messaging service (the view in an MVC world) that provides routes in and out. If it were possible to take a RunRecord once it satisfied the plan and push the RDF to SMW that would be very cool...anyway - back to getting my head around the data model... - Cameron Neylon

What would have to do to get the solubility data on there Egon? - Jean-Claude Bradley

Jean-Claude, I have asked Samuel to reply... this work is his Google Summer of Code project... check his blog -> http://saml.rilspace.org/ - Egon Willighagen

Experimental RDF Import works already, as demoed at http://saml.rilspace.org/even-ni... (Imported snippet at: http://saml.rilspace.org/populat...). Source is available at https://code.google.com/p... I'll be happy to write up an install instruction, but currently things are moving a bit so I think the best if that you request an up-to-date instruction when you need it! - Samuel Lampa

Samuel, perhaps you can set up an ONS demo? Can you import from an existing SPARQL end point? Well, you can use Bioclipse to generate the RDF directly from the Google spreadsheet... Samuel, check these plugins: http://github.com/egonw... - Egon Willighagen

We also have the data in RDF (via Pierre Lindenbaum) if that makes it easier http://anybody.cephb.fr/perso... - Jean-Claude Bradley

Import currently supports RDF/XML, so that is indeed easiest. (Thanks for the link, Will look at the linked RDF!). Additionally the ARC2 library we're using has a number of more parsers, supporting turtle etc, so it should be possible to use these formats as well with a little customization. Egon, does the sparql endpoint support the CONSTRUCT keyword? If so, it should be able to generate the ouputted triples in RDF/XML from the endpoint, no? - Samuel Lampa

I've looked at the dataset now. I need to fix some things to get it working. For example, the <chem:Result> tags create blank nodes (lacking URI), which forces us to either use a randomly created identifier/string as wiki title, or use any of the properties in the data to create it. Any suggestions? Could experiment-id be used? - Samuel Lampa

I also see that I need to improve many things in the import functionality to make it more useful. For example, it would be nice if imported triples were written as mediawiki template calls rather than just semantic mediawiki triple statements, since it would allow to output the data more nicesly, etc. But this seems like a good usecase, so I'd be happy to work on making it useful! - Samuel Lampa

An example can now be seen at http://smwsandbox.rilspace.org/wiki... . I could not import the whole dataset though, since the square brackets in some of the SMILES creates some trouble in the import. Looking into this now. - Samuel Lampa

Ok, so I'm hoping that in the next few days I'll start generating some RDF descriptions as a byproduct of some data analysis. That could start to get rather interesting I think...hassle me if I forget to update you but this will be basically writing out an OREChem Experiment Ontology description of some simple data analysis steps against an EO plan. I need to figure out the best way to write RDF/XML out of Python first and finish the analysis GUI... - Cameron Neylon

Thanks for the link to demo on the wiki Samuel! If you can use XML there is a live feed here http://onsbooks.wikispaces.com/ONS+Sol... The RDF data that I gave you is quite outdated but the XML and spreadsheets are live. - Jean-Claude Bradley

Ok. I should be able to transform the XML feed to the same RDF/XML format as in the linked file, with an XSLT stylesheet. Will look into this next week. - Samuel Lampa

Great Samuel - if that works out we have similar feeds for chemical reaction attempts here http://onsbooks.wikispaces.com/Reactio... - the unique identifier there is the RXID. I would imagine the triples would consist of this compound is a reactant in this reaction; this compound is a product of this reaction; this compound is the solvent of this reaction, etc. - Jean-Claude Bradley

I've managed to transform most of the data now. I could not find data in the XML feed to use for the RDF/XML elements "chem:experiment-id" and "chem:sample-id" though. From where is this data taken? (Also, some additional processing were required, like replacing &-chars with &amp; (making it xml compliant) ... and removing the "xmlns=..." statement without namespace (i.e. missing 'x' in 'xmlns:x'), in the start tag, since it broke the XSLT transformation). - Samuel Lampa

The transformed (RDF/XML) version of the live data is available now at http://smwsandbox.rilspace.org/w... , while the stylesheet is found at http://smwsandbox.rilspace.org/w... (As you can see, data is missing for "chem:experiment-id" and "chem:sample-id"). - Samuel Lampa

Samuel, the basic unit for the solubility data in the XML feed is not an experiment but an average of experimental measurements for a given solvent/solute. Can you use the URL provided as a unique id - like this one http://old.oru.edu/cccda... - Jean-Claude Bradley

Samuel, can you use the ID as the sample-id, e.g. <id>O=C(O)c1cc(ccc1Cl)[N+]([O-])=O.OCCCC</id> and the Provenance as the experiment-id, e.g. <d:Provenance>http://old.oru.edu/cccda...</d:Provenance>? If you need me to move things around or relabel in the feed, let me know. - Andrew Lang

@Jean-Claude Bradley: Ok, I was trying to conform to the RDF/XML format in the selected link, which used experiment-id and sample-id, but if that exact format is not fixed, I could easily change ti, to include things from the XML feed instead. @Andrew Lang: The XML format is fine, I don't think it needs any modificatoin, since XSLT can pick out things anyhow. The question, I guess, is whether to use "experiment-id" and "sample-id" at all in the resulting RDF/XML? - Samuel Lampa

@Samuel. I think matching up as I suggested above would be good. - Andrew Lang

@Andrew Lang: Thanks, the XSLT is now update according to your suggestion. - Samuel Lampa

Unfortunately I will not have more time to work on the XSLT etc, but feel free to use it if you want. I will also try to use this data as example data in order to see that RDF import works well with my extension, so hopefully your data will be fully supported when I'm finished! - Samuel Lampa

(Btw, the extension got a new code repo: http://code.google.com/p... ... and Extension page: http://www.mediawiki.org/wiki...) - Samuel Lampa