Update or synchronize AERS linksets and datasets

[stain]

As pointed out in the support portal by Andrea Splendiani, some of the strange AERS URIs in the loaded AERS dataset don't exist in the corresponding AERS-Drugbank linkset.

Example, from the dataset faers-of-2012-generated-on-2012-07-09.nt:

<http://aers.data2semantics.org/resource/drug/ATARAX-P_%28HYDROXYZINE_HYDROCHLORIDE%29_SYRUP> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://aers.data2semantics.org/vocab/Drug> .
<http://aers.data2semantics.org/resource/drug/ATARAX-P_%28HYDROXYZINE_HYDROCHLORIDE%29_SYRUP> <http://www.w3.org/2000/01/rdf-schema#label> "ATARAX-P (HYDROXYZINE HYDROCHLORIDE) SYRUP" .
<http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://aers.data2semantics.org/vocab/Drug> .
<http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://aers.data2semantics.org/vocab/Drug> .
<http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> <http://www.w3.org/2000/01/rdf-schema#label> "ATARAX-P                           /00058402/" .
<http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> <http://www.w3.org/2000/01/rdf-schema#label> "ATARAX-P                           /00058402/" .
<http://aers.data2semantics.org/resource/involvement/8046146/1018523970> <http://aers.data2semantics.org/vocab/drug> <http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> .
<http://aers.data2semantics.org/resource/involvement/8046146/1018523970> <http://www.w3.org/2000/01/rdf-schema#label> "Involvement of ATARAX-P                           /00058402/ in #8046146" .
<http://aers.data2semantics.org/resource/involvement/8130835/1018840039> <http://aers.data2semantics.org/vocab/drug> <http://aers.data2semantics.org/resource/drug/ATARAX-P___________________________%2F00058402%2F> .
<http://aers.data2semantics.org/resource/involvement/8130835/1018840039> <http://www.w3.org/2000/01/rdf-schema#label> "Involvement of ATARAX-P                           /00058402/ in #8130835" .
<http://aers.data2semantics.org/resource/involvement/8152200/1018970802> <http://aers.data2semantics.org/vocab/drug> <http://aers.data2semantics.org/resource/drug/ATARAX-P_%28HYDROXYZINE_HYDROCHLORIDE%29_SYRUP> .
<http://aers.data2semantics.org/resource/involvement/8152200/1018970802> <http://www.w3.org/2000/01/rdf-schema#label> "Involvement of ATARAX-P (HYDROXYZINE HYDROCHLORIDE) SYRUP in #8152200" .

Yet the linkset do not contain any ATARAX-P, just a few variations of ATARAX:

<http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugs/DB00557> skos:exactMatch <http://aers.data2semantics.org/resource/drug/ATARAX>.
<http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugs/DB00557> skos:exactMatch <http://aers.data2semantics.org/resource/drug/ATARAX_____________________________%2F00058402%2F>.
<http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugs/DB00557> skos:exactMatch <http://aers.data2semantics.org/resource/drug/ATARAXOID>.
<http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugs/DB00557> skos:exactMatch <http://aers.data2semantics.org/resource/drug/ATARAX_____________________________%2F00058403%2F>.
<http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugs/DB00557> skos:exactMatch <http://aers.data2semantics.org/resource/drug/ATARAX_%2F00595201%2F>.

This is probably because the linkset was created 2013-05-06, while the AERS dump faers-of-2012-generated-on-2012-07-09.nt was made 2012-07-09 - and identifiers in AERS probably changed in between.

Note that http://aers.data2semantics.org/ is currently down.

[stain]

Note that the identifiers in AERS (particuarly the linkset) are dubious in themselves, with many duplicates, like:

ZOPICLONE
ZOPICLONE__________________________%28ZOPICLONE
ZOPICLONE_________________________%28ZOPICLONE%29
ZOPICLONE_______________________%28ZOPICLONE%29
ZOPICLONE____________________%28ZOPICLONE%29
ZOPICLONE_____________%28ZOPICLONE%29
ZOPICLONE____________%28ZOPICLONE%29
ZOPICLONE__________%28ZOPICLONE%29
ZOPICLONE_________%28ZOPICLONE%29
ZOPICLONE_____%28ZOPICLONE%29
ZOPICLONE_%28ZOPICLONE%29
ZOPICLONE%28ZOPICLONE%29

Unquoting the %-escaping you get:

ZOPICLONE
ZOPICLONE__________________________(ZOPICLONE
ZOPICLONE_________________________(ZOPICLONE)
ZOPICLONE_______________________(ZOPICLONE)
ZOPICLONE____________________(ZOPICLONE)
ZOPICLONE_____________(ZOPICLONE)
ZOPICLONE____________(ZOPICLONE)
ZOPICLONE__________(ZOPICLONE)
ZOPICLONE_________(ZOPICLONE)
ZOPICLONE_____(ZOPICLONE)
ZOPICLONE_(ZOPICLONE)
ZOPICLONE(ZOPICLONE)

Most of these ___ variants are not used in faers-of-2012-generated-on-2012-07-09.nt - but those that do follow a similar pattern - where the name seem to correspond to a direct transformation of the rdfs:label.

[stain]

This VoID for the linkset proposes the SPARQL query

PREFIX skos: <http://www.w3.org/2004/02/skos/core#> 
CONSTRUCT {
?drugbank_uri skos:closeMatch ?aers_uri
} WHERE {
?drugbank_uri <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www4.wiwiss.fu-berlin.de/drugbank/resource/drugbank/drugs> .
 ?aers_uri skos:closeMatch ?drugbank_uri
}

which I assume was run at http://aers.data2semantics.org/ - those skos:closeMatch relations don't exist in any of the other RDF data.

Perhaps @antonisloizou or @rsiebes have the details on how this linkset should be updated or why the URIs are this weird?

[stain]

Probably delay this fix until 2.1