feat: add pg hybridSearch

spring-ai-core/src/main/java/org/springframework/ai/vectorstore/VectorStore.java

@@ -70,4 +70,16 @@ default List<Document> similaritySearch(String query) {
 		return this.similaritySearch(SearchRequest.query(query));
 	}
 
+	/**
+	 * Retrieves documents by query embedding similarity using the default, If the
+	 * subclass implements this method, use hybrid search, similar search + full text
+	 * search. {@link SearchRequest}'s' search criteria.
+	 * @param request request for set search parameters, such as the query text, topK,
+	 * similarity threshold and metadata filter expressions.
+	 * @return Returns a list of documents that have embeddings similar to the query text
+	 */
+	default List<Document> hybridSearch(SearchRequest request) {
+		return this.similaritySearch(request);
+	}
+
 }

spring-ai-spring-boot-autoconfigure/src/test/java/org/springframework/ai/autoconfigure/vectorstore/pgvector/PgVectorStoreAutoConfigurationIT.java

@@ -95,18 +95,23 @@ public void addAndSearch() {
 
 			vectorStore.add(documents);
 
-			List<Document> results = vectorStore
+			List<Document> similarityResults = vectorStore
 				.similaritySearch(SearchRequest.query("What is Great Depression?").withTopK(1));
+			assertThat(similarityResults).hasSize(1);
+			Document similarityResultDoc = similarityResults.get(0);
+			assertThat(similarityResultDoc.getId()).isEqualTo(documents.get(2).getId());
+			assertThat(similarityResultDoc.getMetadata()).containsKeys("depression", "distance");
 
-			assertThat(results).hasSize(1);
-			Document resultDoc = results.get(0);
-			assertThat(resultDoc.getId()).isEqualTo(documents.get(2).getId());
-			assertThat(resultDoc.getMetadata()).containsKeys("depression", "distance");
+			List<Document> hybridResults = vectorStore
+				.hybridSearch(SearchRequest.query("What is Great Depression?").withTopK(1));
+			assertThat(hybridResults).hasSize(2);
+			Document hybridResultDoc = hybridResults.get(1);
+			assertThat(hybridResultDoc.getMetadata()).containsKeys("depression");
 
 			// Remove all documents from the store
 			vectorStore.delete(documents.stream().map(doc -> doc.getId()).toList());
-			results = vectorStore.similaritySearch(SearchRequest.query("Great Depression").withTopK(1));
-			assertThat(results).hasSize(0);
+			similarityResults = vectorStore.similaritySearch(SearchRequest.query("Great Depression").withTopK(1));
+			assertThat(similarityResults).hasSize(0);
 		});
 	}
 

vector-stores/spring-ai-pgvector-store/src/main/java/org/springframework/ai/vectorstore/PgVectorStore.java

@@ -239,6 +239,21 @@ public List<Document> similaritySearch(SearchRequest request) {
 				new DocumentRowMapper(this.objectMapper), queryEmbedding, queryEmbedding, distance, request.getTopK());
 	}
 
+	@Override
+	public List<Document> hybridSearch(SearchRequest searchRequest) {
+
+		List<Document> similaritySearch = similaritySearch(searchRequest);
+
+		String sql = "SELECT *, ts_rank_cd(to_tsvector(content), plainto_tsquery(?)) AS rank FROM "
+				+ getFullyQualifiedTableName() + " ORDER BY rank DESC LIMIT ?";
+
+		List<Document> keyLikeSearch = this.jdbcTemplate.query(sql, new DocumentRowMapper(this.objectMapper, true),
+				searchRequest.getQuery(), searchRequest.getTopK());
+
+		similaritySearch.addAll(keyLikeSearch);
+		return similaritySearch;
+	}
+
 	public List<Double> embeddingDistance(String query) {
 		return this.jdbcTemplate.query(
 				"SELECT embedding " + this.comparisonOperator() + " ? AS distance FROM " + getFullyQualifiedTableName(),
@@ -425,21 +440,28 @@ private static class DocumentRowMapper implements RowMapper<Document> {
 
 		private ObjectMapper objectMapper;
 
+		private boolean isHybrid = false;
+
 		public DocumentRowMapper(ObjectMapper objectMapper) {
 			this.objectMapper = objectMapper;
 		}
 
+		public DocumentRowMapper(ObjectMapper objectMapper, boolean isHybrid) {
+			this.objectMapper = objectMapper;
+			this.isHybrid = isHybrid;
+		}
+
 		@Override
 		public Document mapRow(ResultSet rs, int rowNum) throws SQLException {
 			String id = rs.getString(COLUMN_ID);
 			String content = rs.getString(COLUMN_CONTENT);
 			PGobject pgMetadata = rs.getObject(COLUMN_METADATA, PGobject.class);
 			PGobject embedding = rs.getObject(COLUMN_EMBEDDING, PGobject.class);
-			Float distance = rs.getFloat(COLUMN_DISTANCE);
-
 			Map<String, Object> metadata = toMap(pgMetadata);
-			metadata.put(COLUMN_DISTANCE, distance);
-
+			if (!isHybrid) {
+				Float distance = rs.getFloat(COLUMN_DISTANCE);
+				metadata.put(COLUMN_DISTANCE, distance);
+			}
 			Document document = new Document(id, content, metadata);
 			document.setEmbedding(toDoubleList(embedding));