Merge pull request #1569 from inception-project/feature/1549-Filter-e…

…ntity-candidates-by-description #1549 - Filter entity candidates by description
inception-project · Dec 18, 2019 · 263480e · 263480e
2 parents a06c1f6 + 9bf550a
commit 263480e
Show file tree

Hide file tree

Showing 2 changed files with 53 additions and 2 deletions.
diff --git a/...-ui-kb/src/main/java/de/tudarmstadt/ukp/inception/ui/kb/feature/ConceptFeatureEditor.java b/...-ui-kb/src/main/java/de/tudarmstadt/ukp/inception/ui/kb/feature/ConceptFeatureEditor.java
@@ -20,13 +20,17 @@
 import static de.tudarmstadt.ukp.clarin.webanno.support.lambda.LambdaBehavior.visibleWhen;
 import static java.util.Arrays.asList;
 import static java.util.Collections.emptyList;
+import static org.apache.commons.lang3.StringUtils.containsIgnoreCase;
 import static org.apache.commons.lang3.StringUtils.isNotBlank;
+import static org.apache.commons.lang3.StringUtils.substringAfter;
+import static org.apache.commons.lang3.StringUtils.substringBefore;
 import static org.apache.wicket.event.Broadcast.BUBBLE;
 import static org.apache.wicket.markup.head.JavaScriptHeaderItem.forReference;
 
 import java.util.Collections;
 import java.util.List;
 import java.util.Locale;
+import java.util.stream.Collectors;
 
 import org.apache.commons.lang3.StringUtils;
 import org.apache.uima.cas.CAS;
@@ -160,6 +164,29 @@ private List<KBHandle> getCandidates(IModel<AnnotatorState> aStateModel,
         if (aInput == null) {
             return emptyList();
         }
+
+        String input = aInput;
+
+        // Extract filter on the description
+        final String descriptionFilter;
+        if (input.contains("::")) {
+            descriptionFilter = substringAfter(input , "::");
+            input = substringBefore(input , "::");
+        }
+        else {
+            descriptionFilter = null;
+        }
+
+        // Extract exact match filter on the query
+        boolean labelFilter = false;
+        String trimmedInput = input.trim();
+        if (trimmedInput.length() > 2 && trimmedInput.startsWith("\"")
+                && trimmedInput.endsWith("\"")) {
+            input = StringUtils.substring(trimmedInput, 1, -1);
+            labelFilter = true;
+        }
+
+        final String finalInput = input;
 
         List<KBHandle> choices;
         try {
@@ -185,8 +212,8 @@ private List<KBHandle> getCandidates(IModel<AnnotatorState> aStateModel,
                     : -1;
 
             choices = clService.getLinkingInstancesInKBScope(traits.getRepositoryId(),
-                    traits.getScope(), traits.getAllowedValueType(), aInput, mention, mentionBegin,
-                    cas, feat.getProject());
+                    traits.getScope(), traits.getAllowedValueType(), finalInput, mention,
+                    mentionBegin, cas, feat.getProject());
         }
         catch (Exception e) {
             choices = asList(new KBHandle("http://ERROR", "ERROR", e.getMessage(), "en"));
@@ -196,6 +223,19 @@ private List<KBHandle> getCandidates(IModel<AnnotatorState> aStateModel,
                 .find(IPartialPageRequestHandler.class)
                 .ifPresent(target -> target.addChildren(getPage(), IFeedback.class));
         }
+
+        if (labelFilter) {
+            choices = choices.stream()
+                    .filter(kb -> containsIgnoreCase(kb.getUiLabel(), finalInput))
+                    .collect(Collectors.toList());
+        }
+
+        if (isNotBlank(descriptionFilter)) {
+            choices = choices.stream()
+                    .filter(kb -> containsIgnoreCase(kb.getDescription(), descriptionFilter))
+                    .collect(Collectors.toList());
+        }
+
         return choices;
     }
 

diff --git a/...ETA-INF/asciidoc/user-guide/annotation_create-annotations_concept-features.adoc b/...ETA-INF/asciidoc/user-guide/annotation_create-annotations_concept-features.adoc
@@ -2,3 +2,14 @@
 
 Concept features are shown as an auto-complete field. Typing into the field triggers a query against
 the knowledge base and displays candidates in a dropdown list.
+
+.Filtering by description
+The query entered into the field only matches against the label of the knowledge base items, not
+against their description. However, you can filter the candidates by their description. E.g. if you
+wish to find all knowledge base items with `Obama` in the label and `president` in the description,
+then you can write `Obama :: president`. A case-insensitive matching is being used.
+
+.Substring matching
+Depending on the knowledge base and full-text mode being used, there may be fuzzy matching. To
+filter the candidate list down to those candidates which contain a particular substring, put
+double quotes around the query, e.g. `"County Carlow"`. A case-insensitive matching is being used.