INSPIRE themes facets are not properly grouped

[etj]

When there are metadata records in multiple languages, the search UI shows different groups for the same theme, one for each language. So for instance, if we have records with theme "Buildings" and some other records with theme "Gebäude", we'll get both labels in the UI.

It is expected that the entries related to a same theme in the different languages should be grouped together, and the theme label should be displayed in the UI current language.

[pvgenuchten]

Hi Emanuele, this is an incidentally re-occurring issue, see for example https://trac.osgeo.org/geonetwork/ticket/326. There are globally 2 ways 2 resolve this (or enlighten me if you see more options):

• In stead of indexing the actual keyword, we can index the uri to the keyword http://inspire.ec.europa.eu/theme/pf/. Before doing a search we should match a search-string (in a certain language) to a keyword-uri and use the uri in the search. And when displaying the facets we should translate the keyword-uri to it's label.
• At indexation we can add for a number of languages the translated keyword to the index, the metadata will then be found using any of those languages, on display time we filter the facets on only those facets that are in the current language.

The first approach seems most appropriate, but it also presents the biggest impact, that's why generally users opt for the second approach.

[josegar74]

The first approach is feasible, ideally requires to use org.fao.geonet.kernel.search.TermUriTranslator in the facet configuration, so the facet values are displayed with the proper text, while filters of the facet use the uri value.

But seem facets in config-summary.xml, can't use the translator classes, unless I'm missing something.

[pvgenuchten]

The current approach for inspiretheme is to index for each referenced theme also a value inspiretheme_en (english translated theme). If you would use that field as a facet (and translate in the UI the english themes to the selected UI language), you would have the desired behaviour. But this approach does not resolve the general keywords from multilingual thesauri issue.

[etj]

I'm going along these lines:

• store the Concept URI into the lucene index:

<xsl:variable name="inspireThemeURI" select="$inspire-theme[skos:prefLabel=$keyword]/@rdf:about"/>
<Field name="inspirethemeuri" string="{$inspireThemeURI}" store="true" index="true"/>

• add the facet (including translator):

<facets>
   <facet name="inspireThemeURI" indexKey="inspirethemeuri" label="inspireThemesURI"/>
   ...
<summaryType name="hits">
   <item facet="inspireThemeURI" sortBy="value" sortOrder="asc" max="35" translator="term:http://geonetwork-opensource.org/inspire-theme"/>
   ...

In order to get the results:

<inspireThemesURI>
   <inspireThemeURI count="10" name="http://rdfdata.eionet.europa.eu/inspirethemes/themes/1" label="Sistemi di coordinate" />
   <inspireThemeURI count="24" name="http://rdfdata.eionet.europa.eu/inspirethemes/themes/10" label="Elevazione" />
   <inspireThemeURI count="15" name="http://rdfdata.eionet.europa.eu/inspirethemes/themes/11" label="Copertura del suolo" />

but in this way we lose the theme acronym, thus we also lose the css styles. Probably some less trick may solve the problem, but the only solution I see at the moment is to duplicate the iti- css rules to also work with the number in the ending part of the URI. Any suggestion?