Multi word text searches cause error in find_nanopubs_with_text

[raar1]

Hi Tobias,

It seems that search terms with more than one word are causing find_nanopubs_with_text to error out. For example, if you try typing "homo sapiens" into the text field on Tapas ( http://grlc.nanopubs.lod.labs.vu.nl/tapas/tapas.html?op=/find_nanopubs_with_text ) then you get the following error dump:

Virtuoso 37000 Error XM029: Free-text expression, line 1: syntax error at sapiens


SPARQL query:
define sql:big-data-const 0 #+ summary: Find nanopublications that contain the given text
#+ method: GET
#+ pagination: 1000
#+ endpoint_in_url: False

prefix np: <http://www.nanopub.org/nschema#>
prefix npa: <http://purl.org/nanopub/admin/>
prefix dct: <http://purl.org/dc/terms/>
prefix luc: <http://www.ontotext.com/owlim/lucene#>   # for GraphDB
#prefix bif: <http://www.openlinksw.com/schemas/bif#>  # for Virtuoso

select ?np ?graphpred ?subj ?pred ?v ?date ?pubkey where {
  graph npa:graph {
    ?np npa:hasHeadGraph ?h .
    ?np npa:creationDay ?__day_iri .
    ?np npa:creationMonth ?__month_iri .
    ?np npa:creationYear ?__year_iri .
    ?np dct:created ?date .
  }
  graph ?h {
    ?np ?__graphpred_iri ?g .
    bind(?__graphpred_iri as ?graphpred)
  }
  graph ?g {
    { ?subj ?pred ?v . ?v luc:npIdx "homo sapiens" . }
    union
    { ?subj ?pred ?v .  ?v <bif:contains> "homo sapiens" . }
  }
  optional {
    graph npa:graph {
      ?np npa:hasValidSignatureForPublicKey ?pubkey .
    }
    # This duplication is a workaround for a Virtuoso bug (otherwise <bif:contains> above overgenerates):
    graph npa:graph {
      ?np npa:hasValidSignatureForPublicKey ?pubkey .
    }
  }
}
 LIMIT 1000 OFFSET 0

However, if you just enter 'homo' then it works as normal. Any way we can fix this query to allow multi-word searches?

[tkuhn]

Multiword searches are supported but you need to use Boolean syntax, i.e. "homo AND sapiens" instead of "homo sapiens". Sorry that this isn't very well documented.

[raar1]

Ah, ok I see. I also notice that if you use single quotes (e.g. 'homo sapiens') then it works as a single search term.