clear webservice facet

[dietuyt]

Right now it is not always clear where to find CLARIN web services. This is a meta-ticket: the solution might be to

• improve the CMDI descriptions
• improve the profiles used (adding the right ISOcat links)
• create extra mappings (from profile id to a facet value e.g.)

Some examples that should be listed as a web service:

• ​Adelheid (Components/ToolService/ApplicationType=web service)

• ​weblicht services

[dietuyt]

​Some records from the LRT inventory are correctly classified as web service already:

using LrtInventoryResource/LrtCommon/ResourceType = Web Service

[herste]

maak lijstje van alle componentprofiles waarvan we de instanties onder tools en services willen zien; componentprofile namen zitten al in de vlo.

[dietuyt]

a list of the webservice profiles (@Herman: can you please stick to English ;-) :

• clarin.eu:cr1:p_1311927752306
• clarin.eu:cr1:p_1320657629644

[herste]

If we list all the Component-types as a facet, would this ticket also count as closed?

[dietuyt]

I would suggest to couple this facet (resourcetype) to:

• ​http://www.isocat.org/datcat/DC-3806 (ResourceClass?, link for LrtInventory? resources)
• ​http://www.isocat.org/datcat/DC-3786 (Application type, link for ToolService? profile)

That would take away the need to map a set of hard-coded profiles to a facet value. It is also more future-proof.

Note to myself:

At least for the LRT inventory the following datcats have been introduced as of 2012-09-06:

<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4396">Multimodal Corpus</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4393">Spoken Corpus</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4394">Written Corpus</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-3988">Aligned Corpus</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-3949">Treebank</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4363">Grammar</item>
<item AppInfo="" ConceptLink="">N-Gram Model</item>
<item AppInfo="" ConceptLink="">Terminological Resource</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4360">Lexicon / Knowledge Source</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-4584">Web Service</item>
<item AppInfo="" ConceptLink="http://www.isocat.org/datcat/DC-2599">other</item>

the following types will be added (also to LRT inventory Resource (field "Resource type") as replacement for LRT inventory Tool):

• ​http://www.isocat.org/datcat/DC-4585 (local/desktop application)
• ​http://www.isocat.org/datcat/DC-4583 (web application)

[dietuyt]

Tübingen has added an Application type field to their profile and will add it to the weblicht instances (mail Thorsten):

Hi Tom,

for interoperability asked for by Dieter I added the element
ApplicationType (with the only possible value webService) as a child
element to the Service element, which itself is a child of the
WebLichtWebService element.

Currently I leave this as an optional element as not to render all
existing weblicht-CMDI-files invalid, but maybe Alex can add the element
to the weblicht service descriptions.

[dietuyt]

Hi Kees Jan, could you check if the mappings mentioned in https://trac.clarin.eu/ticket/195#comment:6 have been configured at the VLO?

[teckart@informatik.uni-leipzig.de]

Added DC-3786 (application type) to facet 'resourceClass' (r5252)

[Twan Goosen]

Replying to teckart@…:

Added DC-3786 (application type) to facet 'resourceClass' (r5252)

Is it safe to add this to maintenance release 3.0.1?

[teckart@informatik.uni-leipzig.de]

I tested the new facet configuration with ~200K CMDI files without any obvious problems. Therefore I would say it can be added to 3.0.1.

[Twan Goosen]

Replying to teckart@…:

I tested the new facet configuration with ~200K CMDI files without any obvious problems. Therefore I would say it can be added to 3.0.1.

Merged to 3.0 branch in r5253

[Twan Goosen]

I think this should be closed, Thomas / Dieter can you confirm that the change applied in [5253] is sufficient?

[teckart@informatik.uni-leipzig.de]

DC-3806 and DC-3786 seem to work quite well.

Before closing the ticket could you check what happened with the Webservices from the DTA (like ​http://catalog.clarin.eu/oai-harvester/clarin/results/cmdi/Deutsches_Textarchiv/WebLichtWebServices_1.xml)? They should have "webService" as resourcetype like the ones from Tübingen or Prague but can't be found in the VLO right now (​http://catalog.clarin.eu/vlo/search?q=SynCoP&fq=resourceClass:webService&fq=collection:WebLicht+Webservice+Orchestrator).
Maybe an import problem?

[herold@bbaw.de]

We will look into the DTA (DWDS) webservice descriptions but this can be closed independently as it is fixed on the VLO's end.