test: add integration tests for multi-vector query, UNF/NOINDEX, and …

core/src/test/java/com/redis/vl/query/MultiVectorQueryIntegrationTest.java

@@ -0,0 +1,276 @@
+package com.redis.vl.query;
+
+import static org.assertj.core.api.Assertions.*;
+
+import com.redis.vl.BaseIntegrationTest;
+import com.redis.vl.index.SearchIndex;
+import com.redis.vl.schema.*;
+import java.util.*;
+import org.junit.jupiter.api.*;
+
+/**
+ * Integration tests for Multi-Vector Query support (#402).
+ *
+ * <p>Tests simultaneous search across multiple vector fields with weighted score combination.
+ *
+ * <p>Python reference: PR #402 - Multi-vector query support
+ */
+@Tag("integration")
+@DisplayName("Multi-Vector Query Integration Tests")
+@TestMethodOrder(MethodOrderer.OrderAnnotation.class)
+class MultiVectorQueryIntegrationTest extends BaseIntegrationTest {
+
+  private static final String INDEX_NAME = "multi_vector_test_idx";
+  private static SearchIndex searchIndex;
+
+  @BeforeAll
+  static void setupIndex() {
+    // Clean up any existing index
+    try {
+      unifiedJedis.ftDropIndex(INDEX_NAME);
+    } catch (Exception e) {
+      // Ignore if index doesn't exist
+    }
+
+    // Create schema with multiple vector fields
+    IndexSchema schema =
+        IndexSchema.builder()
+            .name(INDEX_NAME)
+            .prefix("product:")
+            .field(TextField.builder().name("title").build())
+            .field(TextField.builder().name("description").build())
+            .field(TagField.builder().name("category").build())
+            .field(NumericField.builder().name("price").sortable(true).build())
+            // Text embeddings (3 dimensions)
+            .field(
+                VectorField.builder()
+                    .name("text_embedding")
+                    .dimensions(3)
+                    .distanceMetric(VectorField.DistanceMetric.COSINE)
+                    .build())
+            // Image embeddings (2 dimensions)
+            .field(
+                VectorField.builder()
+                    .name("image_embedding")
+                    .dimensions(2)
+                    .distanceMetric(VectorField.DistanceMetric.COSINE)
+                    .build())
+            .build();
+
+    searchIndex = new SearchIndex(schema, unifiedJedis);
+    searchIndex.create();
+
+    // Insert test documents with multiple vector embeddings
+    Map<String, Object> doc1 = new HashMap<>();
+    doc1.put("id", "1");
+    doc1.put("title", "Red Laptop");
+    doc1.put("description", "Premium laptop");
+    doc1.put("category", "electronics");
+    doc1.put("price", 1200);
+    doc1.put("text_embedding", new float[] {0.1f, 0.2f, 0.3f});
+    doc1.put("image_embedding", new float[] {0.5f, 0.5f});
+
+    Map<String, Object> doc2 = new HashMap<>();
+    doc2.put("id", "2");
+    doc2.put("title", "Blue Phone");
+    doc2.put("description", "Budget smartphone");
+    doc2.put("category", "electronics");
+    doc2.put("price", 300);
+    doc2.put("text_embedding", new float[] {0.4f, 0.5f, 0.6f});
+    doc2.put("image_embedding", new float[] {0.3f, 0.4f});
+
+    Map<String, Object> doc3 = new HashMap<>();
+    doc3.put("id", "3");
+    doc3.put("title", "Green Tablet");
+    doc3.put("description", "Mid-range tablet");
+    doc3.put("category", "electronics");
+    doc3.put("price", 500);
+    doc3.put("text_embedding", new float[] {0.7f, 0.8f, 0.9f});
+    doc3.put("image_embedding", new float[] {0.1f, 0.2f});
+
+    // Load all documents
+    searchIndex.load(Arrays.asList(doc1, doc2, doc3), "id");
+
+    // Wait for indexing
+    try {
+      Thread.sleep(100);
+    } catch (InterruptedException e) {
+      Thread.currentThread().interrupt();
+    }
+  }
+
+  @AfterAll
+  static void cleanupIndex() {
+    if (searchIndex != null) {
+      try {
+        searchIndex.drop();
+      } catch (Exception e) {
+        // Ignore
+      }
+    }
+  }
+
+  @Test
+  @Order(1)
+  @DisplayName("Should create multi-vector query with single vector")
+  void testSingleVectorQuery() {
+    Vector textVec =
+        Vector.builder()
+            .vector(new float[] {0.1f, 0.2f, 0.3f})
+            .fieldName("text_embedding")
+            .dtype("float32")
+            .weight(1.0)
+            .build();
+
+    MultiVectorQuery query = MultiVectorQuery.builder().vector(textVec).numResults(10).build();
+
+    assertThat(query.getVectors()).hasSize(1);
+    assertThat(query.getNumResults()).isEqualTo(10);
+
+    Map<String, Object> params = query.toParams();
+    assertThat(params).containsKey("vector_0");
+    assertThat(params.get("vector_0")).isInstanceOf(byte[].class);
+  }
+
+  @Test
+  @Order(2)
+  @DisplayName("Should create multi-vector query with multiple vectors")
+  void testMultipleVectorsQuery() {
+    Vector textVec =
+        Vector.builder()
+            .vector(new float[] {0.1f, 0.2f, 0.3f})
+            .fieldName("text_embedding")
+            .weight(0.7)
+            .build();
+
+    Vector imageVec =
+        Vector.builder()
+            .vector(new float[] {0.5f, 0.5f})
+            .fieldName("image_embedding")
+            .weight(0.3)
+            .build();
+
+    MultiVectorQuery query =
+        MultiVectorQuery.builder().vectors(textVec, imageVec).numResults(10).build();
+
+    assertThat(query.getVectors()).hasSize(2);
+
+    // Verify params
+    Map<String, Object> params = query.toParams();
+    assertThat(params).containsKeys("vector_0", "vector_1");
+
+    // Verify query string format
+    String queryString = query.toQueryString();
+    assertThat(queryString)
+        .contains("@text_embedding:[VECTOR_RANGE 2.0 $vector_0]")
+        .contains("@image_embedding:[VECTOR_RANGE 2.0 $vector_1]")
+        .contains(" | ");
+
+    // Verify scoring
+    String formula = query.getScoringFormula();
+    assertThat(formula).contains("0.70 * score_0").contains("0.30 * score_1");
+  }
+
+  @Test
+  @Order(3)
+  @DisplayName("Should combine multi-vector query with filter expression")
+  void testMultiVectorQueryWithFilter() {
+    Vector textVec =
+        Vector.builder().vector(new float[] {0.1f, 0.2f, 0.3f}).fieldName("text_embedding").build();
+
+    Filter filter = Filter.tag("category", "electronics");
+
+    MultiVectorQuery query =
+        MultiVectorQuery.builder().vector(textVec).filterExpression(filter).numResults(5).build();
+
+    String queryString = query.toQueryString();
+    assertThat(queryString).contains(" AND ").contains("@category:{electronics}");
+  }
+
+  @Test
+  @Order(4)
+  @DisplayName("Should calculate score from multiple vectors with different weights")
+  void testWeightedScoringCalculation() {
+    Vector v1 =
+        Vector.builder()
+            .vector(new float[] {0.1f, 0.2f, 0.3f})
+            .fieldName("text_embedding")
+            .weight(0.6)
+            .build();
+
+    Vector v2 =
+        Vector.builder()
+            .vector(new float[] {0.5f, 0.5f})
+            .fieldName("image_embedding")
+            .weight(0.4)
+            .build();
+
+    MultiVectorQuery query = MultiVectorQuery.builder().vectors(v1, v2).build();
+
+    // Verify individual score calculations
+    Map<String, String> calculations = query.getScoreCalculations();
+    assertThat(calculations).hasSize(2);
+    assertThat(calculations.get("score_0")).isEqualTo("(2 - distance_0)/2");
+    assertThat(calculations.get("score_1")).isEqualTo("(2 - distance_1)/2");
+
+    // Verify combined scoring formula
+    String formula = query.getScoringFormula();
+    assertThat(formula).isEqualTo("0.60 * score_0 + 0.40 * score_1");
+  }
+
+  @Test
+  @Order(5)
+  @DisplayName("Should support different vector dimensions and dtypes")
+  void testDifferentDimensionsAndDtypes() {
+    Vector v1 =
+        Vector.builder()
+            .vector(new float[] {0.1f, 0.2f, 0.3f}) // 3 dimensions
+            .fieldName("text_embedding")
+            .dtype("float32")
+            .weight(0.5)
+            .build();
+
+    Vector v2 =
+        Vector.builder()
+            .vector(new float[] {0.5f, 0.5f}) // 2 dimensions
+            .fieldName("image_embedding")
+            .dtype("float32")
+            .weight(0.5)
+            .build();
+
+    MultiVectorQuery query = MultiVectorQuery.builder().vectors(v1, v2).build();
+
+    assertThat(query.getVectors().get(0).getVector()).hasSize(3);
+    assertThat(query.getVectors().get(1).getVector()).hasSize(2);
+  }
+
+  @Test
+  @Order(6)
+  @DisplayName("Should specify return fields")
+  void testReturnFields() {
+    Vector textVec =
+        Vector.builder().vector(new float[] {0.1f, 0.2f, 0.3f}).fieldName("text_embedding").build();
+
+    MultiVectorQuery query =
+        MultiVectorQuery.builder()
+            .vector(textVec)
+            .returnFields("title", "price", "category")
+            .build();
+
+    assertThat(query.getReturnFields()).containsExactly("title", "price", "category");
+  }
+
+  @Test
+  @Order(7)
+  @DisplayName("Should use VECTOR_RANGE with threshold 2.0")
+  void testVectorRangeThreshold() {
+    Vector textVec =
+        Vector.builder().vector(new float[] {0.1f, 0.2f, 0.3f}).fieldName("text_embedding").build();
+
+    MultiVectorQuery query = MultiVectorQuery.builder().vector(textVec).build();
+
+    String queryString = query.toQueryString();
+    // Distance threshold hardcoded at 2.0 to include all eligible documents
+    assertThat(queryString).contains("VECTOR_RANGE 2.0");
+  }
+}