Merge pull request #447 from medizininformatik-initiative/feature/446…

…-search-in-original-entry-of-display-if-no-translations-are-available #446 - Search in original entry of display if no translations are available
medizininformatik-initiative · Feb 12, 2025 · 7ec6ecc · 7ec6ecc
2 parents 68ef498 + 7ec2e73
commit 7ec6ecc
Show file tree

Hide file tree

Showing 6 changed files with 43 additions and 14 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -148,7 +148,7 @@ jobs:
     needs: tests
     runs-on: ubuntu-latest
     env:
-      ONTOLOGY_GIT_TAG: v3.0.2-alpha.2
+      ONTOLOGY_GIT_TAG: v3.0.2-alpha.5
       ELASTIC_HOST: http://localhost:9200
       ELASTIC_FILEPATH: https://github.com/medizininformatik-initiative/fhir-ontology-generator/releases/download/TAGPLACEHOLDER/
       ELASTIC_FILENAME: elastic.zip

diff --git a/pom.xml b/pom.xml
@@ -27,7 +27,7 @@
     <java.version>17</java.version>
     <mockwebserver.version>4.12.0</mockwebserver.version>
     <okhttp3.version>4.10.0</okhttp3.version>
-    <ontology-tag>v3.0.2-alpha.2</ontology-tag>
+    <ontology-tag>v3.0.2-alpha.5</ontology-tag>
   </properties>
 
   <dependencies>
@@ -423,7 +423,7 @@
                   <goal>download-single</goal>
                 </goals>
                 <configuration>
-                  <url>https://github.com/medizininformatik-initiative/fhir-ontology-generator/releases/download/${ontology-tag}/</url>
+                  <url>https://github.com/medizininformatik-initiative/fhir-ontology-generator/releases/download/${ontology-tag}</url>
                   <fromFile>backend.zip</fromFile>
                   <toDir>${ontoDirectory.download}</toDir>
                 </configuration>
@@ -435,7 +435,7 @@
                   <goal>download-single</goal>
                 </goals>
                 <configuration>
-                  <url>https://github.com/medizininformatik-initiative/fhir-ontology-generator/releases/download/${ontology-tag}/</url>
+                  <url>https://github.com/medizininformatik-initiative/fhir-ontology-generator/releases/download/${ontology-tag}</url>
                   <fromFile>mapping.zip</fromFile>
                   <toDir>${ontoDirectory.download}</toDir>
                 </configuration>

diff --git a/src/main/java/de/numcodex/feasibility_gui_backend/terminology/es/CodeableConceptService.java b/src/main/java/de/numcodex/feasibility_gui_backend/terminology/es/CodeableConceptService.java
@@ -80,30 +80,51 @@ private SearchHits<CodeableConceptDocument> findByCodeOrDisplay(String keyword,
       });
     }
 
-    BoolQuery boolQuery;
+    BoolQuery outerBoolQuery;
 
     if (keyword.isEmpty()) {
-      boolQuery = new BoolQuery.Builder()
+      outerBoolQuery = new BoolQuery.Builder()
           .filter(filterTerms.isEmpty() ? List.of() : filterTerms)
           .build();
 
     } else {
-      var mmQuery = new MultiMatchQuery.Builder()
+      // First the "upper" part of the query, when translations are present
+      var mustMultiMatchQueryWithTranslations = new MultiMatchQuery.Builder()
           .query(keyword)
           .fields(List.of("display.de", "display.en", "termcode.code^2"))
           .build();
 
-      boolQuery = new BoolQuery.Builder()
-          .must(List.of(mmQuery._toQuery()))
+      var innerBoolQueryMatchTranslations = new BoolQuery.Builder()
+          .must(List.of(mustMultiMatchQueryWithTranslations._toQuery()))
           .filter(filterTerms.isEmpty() ? List.of() : filterTerms)
           .build();
+
+
+      // The "lower" part that will only be considered when the translations are empty
+      var mustMultiMatchQueryWithOriginal = new MultiMatchQuery.Builder()
+          .query(keyword)
+          .fields(List.of("display.original", "termcode.code^2"))
+          .build();
+
+      var innerBoolQueryMatchOriginal = new BoolQuery.Builder()
+          .must(List.of(mustMultiMatchQueryWithOriginal._toQuery()))
+          .filter(filterTerms.isEmpty() ? List.of() : filterTerms)
+          .build();
+
+      // Combine both parts in the top level bool query
+      outerBoolQuery = new BoolQuery.Builder()
+          .should(List.of(innerBoolQueryMatchTranslations._toQuery(), innerBoolQueryMatchOriginal._toQuery()))
+          .minimumShouldMatch("1")
+          .build();
     }
 
     var query = new NativeQueryBuilder()
-        .withQuery(boolQuery._toQuery())
+        .withQuery(outerBoolQuery._toQuery())
         .withPageable(pageRequest)
         .build();
 
+    log.info(query.getQuery().toString());
+
     return operations.search(query, CodeableConceptDocument.class);
   }
 }
diff --git a/src/main/java/de/numcodex/feasibility_gui_backend/terminology/es/TerminologyEsService.java b/src/main/java/de/numcodex/feasibility_gui_backend/terminology/es/TerminologyEsService.java
@@ -156,13 +156,19 @@ private SearchHits<OntologyListItemDocument> findByNameOrTermcode(String keyword
           .filter(filterTerms.isEmpty() ? List.of() : filterTerms)
           .build();
     } else {
-      var mmQuery = new MultiMatchQuery.Builder()
+      var mmQueryWithTranslations = new MultiMatchQuery.Builder()
           .query(keyword)
           .fields(List.of("display.de", "display.en", "termcode^2"))
           .build();
 
+      var mmQueryWithOriginal = new MultiMatchQuery.Builder()
+          .query(keyword)
+          .fields(List.of("display.original", "termcode^2"))
+          .build();
+
       boolQuery = new BoolQuery.Builder()
-          .must(List.of(mmQuery._toQuery()))
+          .should(List.of(mmQueryWithTranslations._toQuery(), mmQueryWithOriginal._toQuery()))
+          .minimumShouldMatch("1")
           .filter(filterTerms.isEmpty() ? List.of() : filterTerms)
           .build();
     }

diff --git a/src/test/resources/de/numcodex/feasibility_gui_backend/terminology/es/codeable_concept.json b/src/test/resources/de/numcodex/feasibility_gui_backend/terminology/es/codeable_concept.json
@@ -86,7 +86,8 @@
         "properties": {
           "original": {
             "type": "text",
-            "index": false
+            "analyzer": "edge_ngram_analyzer_include_punctuation",
+            "search_analyzer": "lowercase_analyzer"
           },
           "de": {
             "type": "text",

diff --git a/src/test/resources/de/numcodex/feasibility_gui_backend/terminology/es/ontology.json b/src/test/resources/de/numcodex/feasibility_gui_backend/terminology/es/ontology.json
@@ -99,7 +99,8 @@
         "properties": {
           "original": {
             "type": "text",
-            "index": false
+            "analyzer": "edge_ngram_analyzer_include_punctuation",
+            "search_analyzer": "lowercase_analyzer"
           },
           "de": {
             "type": "text",