Closes #1498: Integrate NBFC mining with the community reference extr…

…action algorithm

Adding NBFC mining support in communitiesextract.sql. Supplementing integration tests suite with NBFC case.

iis-wf/iis-wf-referenceextraction/src/main/resources/eu/dnetlib/iis/wf/referenceextraction/community/main_sqlite/oozie_app/lib/scripts/communitiesextract.sql

@@ -73,7 +73,14 @@ setschema 'docid,prev,middle,next' select c1, textwindow2s(keywords(filterstopwo
 ), grants where conceptLabel="INSPIRED RIs" and regexprmatches("\bMIS|INSPIRED", context) 
 ) group by docid
 
-
+-- NBFC
+union all
+select  jdict('documentId',docid, 'conceptId', conceptId, 'confidenceLevel', 0.8,'textsnippet',context) as c1 from (
+  select docid, conceptId, conceptLabel, prev||" "||middle||" "||next as context
+      from (
+        setschema 'docid,prev,middle,next' select c1, textwindow2s(keywords(lower(c2)),7,4,3, 'national biodiversity future cent') from pubs where c2 is not null
+        ),grants where conceptLabel="Italian National Biodiversity Future Center"
+       ) group by docid
 union all
 
 -- DARIAH

iis-wf/iis-wf-referenceextraction/src/test/resources/eu/dnetlib/iis/wf/referenceextraction/community/data/concept.json

@@ -61,6 +61,13 @@
       {"name": "suggestedAcknowledgement", "value":""}
    ]
 }
+{
+   "id": "nbfc", 
+   "label": "Italian National Biodiversity Future Center", 
+   "params":[
+      {"name": "suggestedAcknowledgement", "value":""}
+   ]
+}
 
 
 

iis-wf/iis-wf-referenceextraction/src/test/resources/eu/dnetlib/iis/wf/referenceextraction/community/data/document_text.json

@@ -6,3 +6,4 @@
 {"text":"Acknowledgements This work was supported by ANR programs Investissements d'Avenir: MetaboHub, EMBRC-France, France Génomique, Institut Français de Bioinformatique.","id":"50|dedup_wf_001::f6e0a74009879de5c13103f1e4b99827"}
 {"text":"We acknowledge support of this work by the project “INSPIRED” (MIS 5002550), which is implemented under the Action Reinforcement of the Research and Innovation Infrastructure","id":"50|pmid________::1d7ef02818bae06e15ab18bc6a986ca7"}
 {"text":"Acknowledgements This work was supported by RISIS2 (Research Infrastructure for and Innovation Policy Studies)","id":"50|doi_________::38502b081c386672f71750c99160e836"}
+{"text":"Acknowledgements IDC acknowledge the support of NBFC (National Biodiversity Future Center) to Stazione Zoologica Anton Dohrn, funded by the Italian Ministry of University and Research","id":"50|doi_________::ea5362cf59ec4e505c6870164dc6def2"}

iis-wf/iis-wf-referenceextraction/src/test/resources/eu/dnetlib/iis/wf/referenceextraction/community/data/document_to_community.json

@@ -45,4 +45,12 @@
   "conceptId": "risis",
   "confidenceLevel": 0.8,
   "textsnippet": "Acknowledgements This work was supported by <<< RISIS2 >>> (Research Infrastructure for and Innovation Policy Studies)"
-}
+}
+{
+  "documentId": "50|doi_________::ea5362cf59ec4e505c6870164dc6def2",
+  "conceptId": "nbfc",
+  "confidenceLevel": 0.8,
+  "textsnippet": "acknowledgements idc acknowledge the support of nbfc national biodiversity future center to stazione zoologica"
+}
+
+

iis-wf/iis-wf-referenceextraction/src/test/resources/eu/dnetlib/iis/wf/referenceextraction/community/data/report.json

@@ -5,4 +5,5 @@
 {"key":"processing.referenceExtraction.community.references.byrootid.eric", "type":"COUNTER", "value": "1"}
 {"key":"processing.referenceExtraction.community.references.byrootid.fli", "type":"COUNTER", "value": "1"}
 {"key":"processing.referenceExtraction.community.references.byrootid.inspired-ris", "type":"COUNTER", "value": "1"}
+{"key":"processing.referenceExtraction.community.references.byrootid.nbfc", "type":"COUNTER", "value": "1"}
 {"key":"processing.referenceExtraction.community.references.byrootid.risis", "type":"COUNTER", "value": "1"}