Adds new parameters to the elasticsearch inference API for the rerank task type (#5476) (#5519)

* Adds new parameters to the elasticsearch inference rerank API

* Adds unique inference chunking settings for elasticsearch

* Addresses suggestions

(cherry picked from commit a53ef07)

Co-authored-by: kosabogi <[email protected]>
Co-authored-by: Quentin Pradet <[email protected]>

Author: 3 people

specification/inference/_types/CommonTypes.ts

@@ -1322,6 +1322,24 @@ export class ElasticsearchServiceSettings {
    * The maximum value is 32.
    */
   num_threads: integer
+  /**
+   * Available only for the `rerank` task type using the Elastic reranker model.
+   * Controls the strategy used for processing long documents during inference.
+   *
+   * Possible values:
+   * - `truncate` (default): Processes only the beginning of each document.
+   * - `chunk`: Splits long documents into smaller parts (chunks) before inference.
+   *
+   * When `long_document_strategy` is set to `chunk`, Elasticsearch splits each document into smaller parts but still returns a single score per document.
+   * That score reflects the highest relevance score among all chunks.
+   */
+  long_document_strategy?: string
+  /**
+   * Only for the `rerank` task type.
+   * Limits the number of chunks per document that are sent for inference when chunking is enabled.
+   * If not set, all chunks generated for the document are processed.
+   */
+  max_chunks_per_doc?: integer
 }
 
 export class ElasticsearchTaskSettings {

specification/inference/_types/Services.ts

@@ -50,7 +50,9 @@ import {
  */
 export class InferenceEndpoint {
   /**
-   * Chunking configuration object
+   * The chunking configuration object.
+   * Applies only to the `sparse_embedding` and `text_embedding` task types.
+   * Not applicable to the `rerank`, `completion`, or `chat_completion` task types.
    */
   chunking_settings?: InferenceChunkingSettings
   /**

specification/inference/put_elasticsearch/PutElasticsearchRequest.ts

@@ -79,6 +79,8 @@ export interface Request extends RequestBase {
   body: {
     /**
      * The chunking configuration object.
+     * Applies only to the `sparse_embedding` and `text_embedding` task types.
+     * Not applicable to the `rerank`, `completion`, or `chat_completion` task types.
      * @ext_doc_id inference-chunking
      */
     chunking_settings?: InferenceChunkingSettings