add context and unit parameters to ESOR concept lookup

DataONEorg · Apr 5, 2016 · 3d078a1 · 3d078a1
1 parent 0ad6920
commit 3d078a1
Show file tree

Hide file tree

Showing 11 changed files with 46 additions and 26 deletions.
diff --git a/src/main/java/org/dataone/annotator/generator/AnnotationGenerator.java b/src/main/java/org/dataone/annotator/generator/AnnotationGenerator.java
@@ -47,7 +47,7 @@ public static void initializeCache() {
 			OntDocumentManager.getInstance().addModel(dcterms, ModelFactory.createOntologyModel().read(dcterms_source));
 			OntDocumentManager.getInstance().addModel(foaf, ModelFactory.createOntologyModel().read(foaf_source));
 			OntDocumentManager.getInstance().addModel(prov, ModelFactory.createOntologyModel().read(prov));
-			OntDocumentManager.getInstance().addModel(cito, ModelFactory.createOntologyModel().read(cito));
+			//OntDocumentManager.getInstance().addModel(cito, ModelFactory.createOntologyModel().read(cito));
 			cacheInitialized = true;
 		}
 	}

diff --git a/src/main/java/org/dataone/annotator/generator/json/JsonAnnotatorGenerator.java b/src/main/java/org/dataone/annotator/generator/json/JsonAnnotatorGenerator.java
@@ -120,6 +120,9 @@ private Map<Identifier, String> generateAnnotationsFromEML(Identifier metadataPi
     	DataPackage dataPackage = this.getDataPackage(metadataPid);
     	SystemMetadata sysMeta = D1Client.getCN().getSystemMetadata(null, metadataPid);
 
+    	// TODO: use abstract content for context
+    	String context = dataPackage.getTitle();
+
 		Map<Identifier, String> annotations = new HashMap<Identifier, String>();
 
 		// loop through the tables and attributes
@@ -184,7 +187,7 @@ private Map<Identifier, String> generateAnnotationsFromEML(Identifier metadataPi
 
 
 						// look up concepts for all the attribute text we have
-						List<ConceptItem> concepts = conceptMatcher.getConcepts(attributeText.toString());
+						List<ConceptItem> concepts = conceptMatcher.getConcepts(attributeText.toString(), attributeUnit, context);
 						if (concepts != null && concepts.size() > 0) {
 
 							// debug
@@ -238,7 +241,7 @@ private Map<Identifier, String> generateAnnotationsFromEML(Identifier metadataPi
 
 				// use an orcid if we can find one from their system
 				String creatorText = creators.get(0).getOrganization() + " " + creators.get(0).getSurName() + " " + creators.get(0).getGivenNames();
-				List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText);
+				List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText, null, null);
 				if (concepts != null) {
 					JSONObject annotation = createAnnotationTemplate(sysMeta);
 					JSONArray creatorTags = new JSONArray();
@@ -370,7 +373,9 @@ private Map<Identifier, String> generateAnnotationsFromIndex(Identifier metadata
         }
 
     	SystemMetadata sysMeta = D1Client.getCN().getSystemMetadata(null, metadataPid);
-
+
+    	String context = solrDoc.get("abstract").toString();
+
 		Map<Identifier, String> annotations = new HashMap<Identifier, String>();
 
 		// loop through the attributes
@@ -423,7 +428,7 @@ private Map<Identifier, String> generateAnnotationsFromIndex(Identifier metadata
 				annotation.put("ranges", ranges);
 
 				// look up concepts for all the attribute text we have
-				List<ConceptItem> concepts = conceptMatcher.getConcepts(attributeName);
+				List<ConceptItem> concepts = conceptMatcher.getConcepts(attributeName, null, context);
 				if (concepts != null && concepts.size() > 0) {
 					// add the tags
 					JSONArray tags = new JSONArray();
@@ -462,7 +467,7 @@ private Map<Identifier, String> generateAnnotationsFromIndex(Identifier metadata
 				String creator = creators.get(0).toString();
 				// use an orcid if we can find one from their system
 				String creatorText = creator;
-				List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText);
+				List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText, null, null);
 				if (concepts != null && concepts.size() > 0) {
 					JSONObject annotation = createAnnotationTemplate(sysMeta);
 					JSONArray creatorTags = new JSONArray();

diff --git a/src/main/java/org/dataone/annotator/generator/oa/OAAnnotationGenerator.java b/src/main/java/org/dataone/annotator/generator/oa/OAAnnotationGenerator.java
@@ -135,7 +135,7 @@ public Map<Identifier, String> generateAnnotations(Identifier metadataPid) throw
 		if (creators != null && creators.size() > 0) {	
 			// use an orcid if we can find one from their system
 			String creatorText = creators.get(0).getOrganization() + " " + creators.get(0).getSurName() + " " + creators.get(0).getGivenNames();
-			List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText);
+			List<ConceptItem> concepts = orcidMatcher.getConcepts(creatorText, null, null);
 			if (concepts != null) {
 				String orcidUri = concepts.get(0).getUri().toString();
 				p1 = m.createIndividual(orcidUri, personClass);
@@ -259,7 +259,7 @@ private Resource lookupStandard(OntClass standardClass, Attribute attribute) thr
 		String unit = attribute.getUnit().toLowerCase();
 
 		// look up the concept using the matcher
-		List<ConceptItem> concepts = conceptMatcher.getConcepts(unit);
+		List<ConceptItem> concepts = conceptMatcher.getConcepts(unit, null, null);
 		return ResourceFactory.createResource(concepts.get(0).getUri().toString());
 		//return BioPortalService.lookupAnnotationClass(standardClass, unit, OBOE_SBC);
 	}
@@ -271,7 +271,7 @@ private Resource lookupCharacteristic(OntClass characteristicClass, Attribute at
 		String text = label + " " + definition;
 
 		// look up the concept using the matcher
-		List<ConceptItem> concepts = conceptMatcher.getConcepts(text);
+		List<ConceptItem> concepts = conceptMatcher.getConcepts(text, null, null);
 		return ResourceFactory.createResource(concepts.get(0).getUri().toString());
 		//return BioPortalService.lookupAnnotationClass(characteristicClass, text, OBOE_SBC);
 
@@ -283,7 +283,7 @@ private Resource lookupEntity(OntClass entityClass, Entity entity) throws Except
 		String definition = entity.getDefinition();
 
 		// look up the concept using the matcher
-		List<ConceptItem> concepts = conceptMatcher.getConcepts(definition);
+		List<ConceptItem> concepts = conceptMatcher.getConcepts(definition, null, null);
 		return ResourceFactory.createResource(concepts.get(0).getUri().toString());
 		//return BioPortalService.lookupAnnotationClass(entityClass, definition, OBOE_SBC);
 

diff --git a/src/main/java/org/dataone/annotator/matcher/ConceptMatcher.java b/src/main/java/org/dataone/annotator/matcher/ConceptMatcher.java
@@ -5,7 +5,7 @@
 
 public interface ConceptMatcher {
 
-	public List<ConceptItem> getConcepts(String fullText) throws Exception;
+	public List<ConceptItem> getConcepts(String fullText, String unit, String context) throws Exception;
 
 	public List<ConceptItem> getConcepts(Map<String, String> terms) throws Exception;
 

diff --git a/src/main/java/org/dataone/annotator/matcher/ManualConceptMatcher.java b/src/main/java/org/dataone/annotator/matcher/ManualConceptMatcher.java
@@ -11,7 +11,7 @@
 public class ManualConceptMatcher implements ConceptMatcher {
 
 	@Override
-	public List<ConceptItem> getConcepts(String fullText) throws Exception {
+	public List<ConceptItem> getConcepts(String fullText, String unit, String context) throws Exception {
 		// TODO Auto-generated method stub
 		return null;
 	}

diff --git a/src/main/java/org/dataone/annotator/matcher/bioportal/BioPortalService.java b/src/main/java/org/dataone/annotator/matcher/bioportal/BioPortalService.java
@@ -45,7 +45,7 @@ public BioPortalService() {
 
 
     @Override
-    public List<ConceptItem> getConcepts(String text) throws Exception {
+    public List<ConceptItem> getConcepts(String text, String unit, String context) throws Exception {
 
     	// limit suggested annotations to MeasurementType subclasses.
     	//OntClass measurementTypeClass = mtg.getMeasurementTypeClass();
@@ -62,7 +62,7 @@ public List<ConceptItem> getConcepts(Map<String, String> queryItems) throws Exce
 			sb.append(value);
 			sb.append(" ");
 		}
-		return getConcepts(sb.toString());
+		return getConcepts(sb.toString(), null, null);
 	}
 
     /**

diff --git a/src/main/java/org/dataone/annotator/matcher/esor/EsorClient.java b/src/main/java/org/dataone/annotator/matcher/esor/EsorClient.java
@@ -13,7 +13,7 @@ public class EsorClient {
     public static void main(String[] args) throws Exception{
         EsorService esorS = new EsorService();
 
-        List<ConceptItem> res = esorS.getConcepts("Litterfall");
+        List<ConceptItem> res = esorS.getConcepts("Litterfall", null, null);
         //List<ConceptItem> res = esorS.getConcepts("carbon%20mass");
         //List<ConceptItem> res = esorS.getConcepts("carbon,mass");
 

diff --git a/src/main/java/org/dataone/annotator/matcher/esor/EsorService.java b/src/main/java/org/dataone/annotator/matcher/esor/EsorService.java
@@ -11,9 +11,12 @@
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
 import org.apache.http.HttpResponse;
+import org.apache.http.NameValuePair;
 import org.apache.http.client.HttpClient;
 import org.apache.http.client.methods.HttpGet;
+import org.apache.http.client.methods.HttpPost;
 import org.apache.http.impl.client.HttpClients;
+import org.apache.http.message.BasicNameValuePair;
 import org.dataone.annotator.matcher.ConceptItem;
 import org.dataone.annotator.matcher.ConceptMatcher;
 import org.dataone.annotator.matcher.QueryItem;
@@ -38,17 +41,17 @@ public EsorService() {
 
 
 	@Override
-	public List<ConceptItem> getConcepts(String fullText) throws Exception {
+	public List<ConceptItem> getConcepts(String fullText, String unit, String context) throws Exception {
 		//merge two results with different escape condition
 		String query = parseTerm(fullText);
 		String escapedSpaceQuery = escapeToSpace(query);
 		String escapedCommaQuery = escapeToComma(query);
 
 		if(true){
-			return lookupEsor(escapedSpaceQuery);
+			return lookupEsor(escapedSpaceQuery, unit, context);
 		}else{
-			List<ConceptItem> res_escapedSpace = lookupEsor(escapedSpaceQuery);
-			List<ConceptItem> res_escapedComma = lookupEsor(escapedCommaQuery);
+			List<ConceptItem> res_escapedSpace = lookupEsor(escapedSpaceQuery, unit, context);
+			List<ConceptItem> res_escapedComma = lookupEsor(escapedCommaQuery, unit, context);
 			return mergeRes(res_escapedSpace, res_escapedComma);
 		}
 
@@ -63,10 +66,10 @@ public List<ConceptItem> getConcepts(Map<String, String> queryItems) throws Exce
 			sb.append(" ");
 		}
 		//return lookupEsor(sb.toString());
-		return getConcepts(sb.toString());
+		return getConcepts(sb.toString(), null, null);
 	}
 
-	private List<ConceptItem> lookupEsor(String query) throws Exception  {
+	private List<ConceptItem> lookupEsor(String query, String unit, String context) throws Exception  {
 
 		HttpClient client = HttpClients.createDefault();
 		// remove quotes for now
@@ -75,11 +78,23 @@ private List<ConceptItem> lookupEsor(String query) throws Exception  {
 		//String uriStr = REST_URL + "?query=" + URLEncoder.encode(query, "UTF-8");
 		String uriStr = REST_URL;
 		//uriStr += "?minScore=2&numResult=10";
-		uriStr +=  "?query=" + query;
+		//uriStr +=  "?query=" + query;
 		log.debug("uriStr=" + uriStr);
 
-		HttpGet method = new HttpGet(uriStr);
+		// use post to handle potentially long parameter values
+		HttpPost method = new HttpPost(uriStr);
 		method.setHeader("Accept", "*/*");
+		ArrayList<NameValuePair> postParameters = new ArrayList<NameValuePair>();
+	    postParameters.add(new BasicNameValuePair("query", query));
+	    if (context != null) {
+		    postParameters.add(new BasicNameValuePair("context", context));
+	    }
+	    if (unit != null) {
+		    postParameters.add(new BasicNameValuePair("unit", unit));
+	    }
+	    //postParameters.add(new BasicNameValuePair("minScore", "2"));
+	    //postParameters.add(new BasicNameValuePair("numResult", "10"));
+
 		HttpResponse response = client.execute(method);
 		int code = response.getStatusLine().getStatusCode();
 		if (2 != code / 100) {

diff --git a/src/main/java/org/dataone/annotator/matcher/orcid/OrcidService.java b/src/main/java/org/dataone/annotator/matcher/orcid/OrcidService.java
@@ -32,7 +32,7 @@ class OrcidService implements ConceptMatcher {
     private static final String REST_URL = "http://pub.orcid.org/v1.1/search/orcid-bio";
 
     @Override
-    public List<ConceptItem> getConcepts(String text) throws Exception {
+    public List<ConceptItem> getConcepts(String text, String unit, String context) throws Exception {
     	return lookupOrcid(text, null, null, null);
 
     }

diff --git a/src/test/java/org/dataone/annotator/matcher/bioportal/BioPortalServiceTest.java b/src/test/java/org/dataone/annotator/matcher/bioportal/BioPortalServiceTest.java
@@ -67,7 +67,7 @@ public void testLookup() {
 			String text = "carbon flux";
 			BioPortalService service = new BioPortalService();
 			List<ConceptItem> results;
-			results = service.getConcepts(text);
+			results = service.getConcepts(text, null, null);
 			String retConcept = results.get(0).getUri().toString();
 			assertEquals("http://purl.dataone.org/odo/ECSO_00000011", retConcept);
 		} catch (Exception e) {

diff --git a/src/test/java/org/dataone/annotator/matcher/esor/EsorServiceTest.java b/src/test/java/org/dataone/annotator/matcher/esor/EsorServiceTest.java
@@ -68,7 +68,7 @@ public void testLookup() {
 			String text = "carbon flux";
 			EsorService service = new EsorService();
 			List<ConceptItem> results;
-			results = service.getConcepts(text);
+			results = service.getConcepts(text, null, null);
 			String retConcept = results.get(0).getUri().toString();
 			System.out.println("retConcept=" + retConcept);
 			assertEquals("http://purl.dataone.org/odo/ECSO_00000011", retConcept);