Add batch_size param for text_embedding processor (#1298) (#1312)

* Add batchSize parameter for text_embedding processor * throw IllegalArgumentException when batchSize is not a positive integer * test: add test cases for BatchSize param * test: exception when batchSize is zero or negative integer * refactor: use assertNotNull for readability & convention * update CHANGELOG about #1298 PR * apply code convention to keep the codes spotless --------- (cherry picked from commit 6c3e68f) Signed-off-by: YeonghyeonKO <[email protected]> Signed-off-by: Thomas Farr <[email protected]> Signed-off-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com> Co-authored-by: Thomas Farr <[email protected]>
opensearch-project · Nov 18, 2024 · 99d0653 · 99d0653
1 parent 810f4bc
commit 99d0653
Show file tree

Hide file tree

Showing 3 changed files with 70 additions and 7 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,6 +4,7 @@ Inspired from [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
 ## [Unreleased 2.x]
 ### Added
 - Added support for disabling typed keys serialization ([#1296](https://github.com/opensearch-project/opensearch-java/pull/1296))
+- Added support for the `batch_size` param on the `text_embedding` processor ([#1298](https://github.com/opensearch-project/opensearch-java/pull/1298))
 
 ### Dependencies
 

diff --git a/...-client/src/main/java/org/opensearch/client/opensearch/ingest/TextEmbeddingProcessor.java b/...-client/src/main/java/org/opensearch/client/opensearch/ingest/TextEmbeddingProcessor.java
@@ -31,6 +31,9 @@ public class TextEmbeddingProcessor extends ProcessorBase implements ProcessorVa
     @Nullable
     private final String description;
 
+    @Nullable
+    private final Integer batchSize;
+
     // ---------------------------------------------------------------------------------------------
 
     private TextEmbeddingProcessor(Builder builder) {
@@ -39,7 +42,7 @@ private TextEmbeddingProcessor(Builder builder) {
         this.modelId = ApiTypeHelper.requireNonNull(builder.modelId, this, "modelId");
         this.fieldMap = ApiTypeHelper.unmodifiableRequired(builder.fieldMap, this, "fieldMap");
         this.description = builder.description;
-
+        this.batchSize = builder.batchSize;
     }
 
     public static TextEmbeddingProcessor of(Function<Builder, ObjectBuilder<TextEmbeddingProcessor>> fn) {
@@ -76,6 +79,14 @@ public final String description() {
         return this.description;
     }
 
+    /**
+     * API name: {@code batch_size}
+     */
+    @Nullable
+    public final Integer batchSize() {
+        return this.batchSize;
+    }
+
     protected void serializeInternal(JsonGenerator generator, JsonpMapper mapper) {
 
         super.serializeInternal(generator, mapper);
@@ -96,7 +107,10 @@ protected void serializeInternal(JsonGenerator generator, JsonpMapper mapper) {
             generator.writeKey("description");
             generator.write(this.description);
         }
-
+        if (this.batchSize != null) {
+            generator.writeKey("batch_size");
+            generator.write(this.batchSize);
+        }
     }
 
     // ---------------------------------------------------------------------------------------------
@@ -114,6 +128,9 @@ public static class Builder extends ProcessorBase.AbstractBuilder<Builder> imple
         @Nullable
         private String description;
 
+        @Nullable
+        private Integer batchSize;
+
         /**
          * Required - API name: {@code model_id}
          */
@@ -150,6 +167,17 @@ public final Builder description(@Nullable String value) {
             return this;
         }
 
+        /**
+         * API name: {@code batch_size}
+         */
+        public final Builder batchSize(@Nullable Integer value) {
+            if (value != null && value <= 0) {
+                throw new IllegalArgumentException("batchSize must be a positive integer");
+            }
+            this.batchSize = value;
+            return this;
+        }
+
         @Override
         protected Builder self() {
             return this;
@@ -183,6 +211,7 @@ protected static void setupTextEmbeddingProcessorDeserializer(ObjectDeserializer
         op.add(Builder::modelId, JsonpDeserializer.stringDeserializer(), "model_id");
         op.add(Builder::fieldMap, JsonpDeserializer.stringMapDeserializer(JsonpDeserializer.stringDeserializer()), "field_map");
         op.add(Builder::description, JsonpDeserializer.stringDeserializer(), "description");
+        op.add(Builder::batchSize, JsonpDeserializer.integerDeserializer(), "batch_size");
     }
 
 }
diff --git a/...ent/src/test/java/org/opensearch/client/opensearch/ingest/TextEmbeddingProcessorTest.java b/...ent/src/test/java/org/opensearch/client/opensearch/ingest/TextEmbeddingProcessorTest.java
@@ -25,30 +25,63 @@ private static TextEmbeddingProcessor.Builder baseTextEmbeddingProcessor() {
     }
 
     @Test
-    public void testJsonRoundtripWithDescription() {
+    public void testJsonRoundtripWithDescriptionAndBatchSize() {
         Processor processor = new Processor.Builder().textEmbedding(
-            baseTextEmbeddingProcessor().description("processor-description").build()
+            baseTextEmbeddingProcessor().description("processor-description").batchSize(1).build()
         ).build();
         String json =
-            "{\"text_embedding\":{\"tag\":\"some-tag\",\"model_id\":\"modelId\",\"field_map\":{\"input_field\":\"vector_field\"},\"description\":\"processor-description\"}}";
+            "{\"text_embedding\":{\"tag\":\"some-tag\",\"model_id\":\"modelId\",\"field_map\":{\"input_field\":\"vector_field\"},\"description\":\"processor-description\",\"batch_size\":1}}";
         TextEmbeddingProcessor deserialized = checkJsonRoundtrip(processor, json).textEmbedding();
 
         assertEquals("modelId", deserialized.modelId());
         assertEquals(baseFieldMap, deserialized.fieldMap());
         assertEquals("processor-description", deserialized.description());
         assertEquals("some-tag", deserialized.tag());
+        assertNotNull(deserialized.batchSize());
+        assertEquals(1, deserialized.batchSize().intValue());
     }
 
     @Test
     public void testJsonRoundtripWithoutDescription() {
-        Processor processor = new Processor.Builder().textEmbedding(baseTextEmbeddingProcessor().build()).build();
+        Processor processor = new Processor.Builder().textEmbedding(baseTextEmbeddingProcessor().batchSize(1).build()).build();
         String json =
-            "{\"text_embedding\":{\"tag\":\"some-tag\",\"model_id\":\"modelId\",\"field_map\":{\"input_field\":\"vector_field\"}}}";
+            "{\"text_embedding\":{\"tag\":\"some-tag\",\"model_id\":\"modelId\",\"field_map\":{\"input_field\":\"vector_field\"},\"batch_size\":1}}";
         TextEmbeddingProcessor deserialized = checkJsonRoundtrip(processor, json).textEmbedding();
 
         assertEquals("modelId", deserialized.modelId());
         assertEquals(baseFieldMap, deserialized.fieldMap());
         assertNull(deserialized.description());
         assertEquals("some-tag", deserialized.tag());
+        assertNotNull(deserialized.batchSize());
+        assertEquals(1, deserialized.batchSize().intValue());
+    }
+
+    @Test
+    public void testJsonRoundtripWithoutBatchSize() {
+        Processor processor = new Processor.Builder().textEmbedding(
+            baseTextEmbeddingProcessor().description("processor-description").build()
+        ).build();
+        String json =
+            "{\"text_embedding\":{\"tag\":\"some-tag\",\"model_id\":\"modelId\",\"field_map\":{\"input_field\":\"vector_field\"},\"description\":\"processor-description\"}}";
+        TextEmbeddingProcessor deserialized = checkJsonRoundtrip(processor, json).textEmbedding();
+
+        assertEquals("modelId", deserialized.modelId());
+        assertEquals(baseFieldMap, deserialized.fieldMap());
+        assertEquals("processor-description", deserialized.description());
+        assertEquals("some-tag", deserialized.tag());
+        assertNull(deserialized.batchSize());
+    }
+
+    @Test
+    public void testInvalidBatchSizeThrowsException() {
+        IllegalArgumentException exceptionWhenBatchSizeIsZero = assertThrows(IllegalArgumentException.class, () -> {
+            new Processor.Builder().textEmbedding(baseTextEmbeddingProcessor().batchSize(0).build()).build();
+        });
+        assertEquals("batchSize must be a positive integer", exceptionWhenBatchSizeIsZero.getMessage());
+
+        IllegalArgumentException exceptionWhenBatchSizeIsNegative = assertThrows(IllegalArgumentException.class, () -> {
+            new Processor.Builder().textEmbedding(baseTextEmbeddingProcessor().batchSize(-1).build()).build();
+        });
+        assertEquals("batchSize must be a positive integer", exceptionWhenBatchSizeIsNegative.getMessage());
     }
 }