GH-3896: Similarity searches with the MariaDBVectorStore do not provide a score

Fixes #3896 Fixes GH-3896 (#3896) In addition to provide the distance as part of the metadata, MariaDBVectorStore calculates the score from the distance and sets via the builder. Signed-off-by: Tilman Holube <tilman@holube.de> Restore distance as part of metadata Signed-off-by: Soby Chacko <soby.chacko@broadcom.com> (cherry picked from commit 891ef2a)

Author: Til7701

vector-stores/spring-ai-mariadb-store/src/main/java/org/springframework/ai/vectorstore/mariadb/MariaDBVectorStore.java

@@ -487,7 +487,13 @@ public Document mapRow(ResultSet rs, int rowNum) throws SQLException {
 
 metadata.put("distance", distance);
 
-return new Document(id, content, metadata);
+// @formatter:off
+return Document.builder()
+.id(id)
+.text(content)
+.metadata(metadata)
+.score(1.0 - distance)
+.build(); // @formatter:on
 }
 
 private Map<String, Object> toMap(String source) {

vector-stores/spring-ai-mariadb-store/src/test/java/org/springframework/ai/vectorstore/mariadb/MariaDBStoreIT.java

@@ -122,20 +122,20 @@ static Stream<Arguments> provideFilters() {
 );
 }
 
-private static boolean isSortedByDistance(List<Document> docs) {
+private static boolean isSortedByScore(List<Document> docs) {
 
-List<Float> distances = docs.stream().map(doc -> (Float) doc.getMetadata().get("distance")).toList();
+List<Double> scores = docs.stream().map(Document::getScore).toList();
 
-if (CollectionUtils.isEmpty(distances) || distances.size() == 1) {
+if (CollectionUtils.isEmpty(scores) || scores.size() == 1) {
 return true;
 }
 
-Iterator<Float> iter = distances.iterator();
-Float current;
-Float previous = iter.next();
+Iterator<Double> iter = scores.iterator();
+Double current;
+Double previous = iter.next();
 while (iter.hasNext()) {
 current = iter.next();
-if (previous > current) {
+if (previous < current) {
 return false;
 }
 previous = current;
@@ -166,7 +166,8 @@ public void addAndSearch(String distanceType) {
 assertThat(results).hasSize(1);
 Document resultDoc = results.get(0);
 assertThat(resultDoc.getId()).isEqualTo(this.documents.get(2).getId());
-assertThat(resultDoc.getMetadata()).containsKeys("meta2", "distance");
+assertThat(resultDoc.getMetadata()).containsKeys("meta2");
+assertThat(resultDoc.getScore()).isBetween(0.0, 1.0);
 
 // Remove all documents from the store
 vectorStore.delete(this.documents.stream().map(doc -> doc.getId()).toList());
@@ -315,7 +316,8 @@ public void documentUpdate(String distanceType) {
 Document resultDoc = results.get(0);
 assertThat(resultDoc.getId()).isEqualTo(document.getId());
 assertThat(resultDoc.getText()).isEqualTo("Spring AI rocks!!");
-assertThat(resultDoc.getMetadata()).containsKeys("meta1", "distance");
+assertThat(resultDoc.getMetadata()).containsKeys("meta1");
+assertThat(resultDoc.getScore()).isBetween(0.0, 1.0);
 
 Document sameIdDocument = new Document(document.getId(),
 "The World is Big and Salvation Lurks Around the Corner",
@@ -329,7 +331,8 @@ public void documentUpdate(String distanceType) {
 resultDoc = results.get(0);
 assertThat(resultDoc.getId()).isEqualTo(document.getId());
 assertThat(resultDoc.getText()).isEqualTo("The World is Big and Salvation Lurks Around the Corner");
-assertThat(resultDoc.getMetadata()).containsKeys("meta2", "distance");
+assertThat(resultDoc.getMetadata()).containsKeys("meta2");
+assertThat(resultDoc.getScore()).isBetween(0.0, 1.0);
 
 dropTable(context);
 });
@@ -350,19 +353,14 @@ public void searchWithThreshold(String distanceType) {
 
 assertThat(fullResult).hasSize(3);
 
-assertThat(isSortedByDistance(fullResult)).isTrue();
+assertThat(isSortedByScore(fullResult)).isTrue();
 
-List<Float> distances = fullResult.stream()
-.map(doc -> (Float) doc.getMetadata().get("distance"))
-.toList();
+List<Double> scores = fullResult.stream().map(Document::getScore).toList();
 
-float threshold = (distances.get(0) + distances.get(1)) / 2;
+double threshold = (scores.get(0) + scores.get(1)) / 2;
 
-List<Document> results = vectorStore.similaritySearch(SearchRequest.builder()
-.query("Time Shelter")
-.topK(5)
-.similarityThreshold(1 - threshold)
-.build());
+List<Document> results = vectorStore.similaritySearch(
+SearchRequest.builder().query("Time Shelter").topK(5).similarityThreshold(threshold).build());
 
 assertThat(results).hasSize(1);
 Document resultDoc = results.get(0);