address review

jimczi · jimczi · commit f837041c6244 · 2025-09-15T22:15:03.000+01:00
diff --git a/modules/reindex/src/main/java/org/elasticsearch/reindex/AbstractAsyncBulkByScrollAction.java b/modules/reindex/src/main/java/org/elasticsearch/reindex/AbstractAsyncBulkByScrollAction.java
@@ -214,7 +214,7 @@ static <Request extends AbstractBulkByScrollRequest<Request>> SearchRequest prep
             // always include vectors in the response unless explicitly set
             var fetchSource = sourceBuilder.fetchSource();
             if (fetchSource == null) {
-                sourceBuilder.fetchSource(FetchSourceContext.of(true, false, new String[0], new String[0]));
+                sourceBuilder.fetchSource(FetchSourceContext.FETCH_ALL_SOURCE_EXCLUDE_INFERENCE_FIELDS);
             } else if (fetchSource.excludeVectors() == null) {
                 sourceBuilder.excludeVectors(false);
             }
diff --git a/x-pack/plugin/inference/src/yamlRestTest/resources/rest-api-spec/test/inference/30_semantic_text_inference.yml b/x-pack/plugin/inference/src/yamlRestTest/resources/rest-api-spec/test/inference/30_semantic_text_inference.yml
@@ -348,6 +348,11 @@ setup:
       indices.create:
         index: destination-index
         body:
+          settings:
+            index:
+              mapping:
+                semantic_text:
+                  use_legacy_format: false
           mappings:
             properties:
               sparse_field:
@@ -370,6 +375,105 @@ setup:
           dest:
             index: destination-index
 
+  - match: { created: 1 }
+  - match: { updated: 0 }
+  - match: { version_conflicts: 0 }
+  - match: { failures: [ ] }
+
+  - do:
+      indices.refresh: { }
+
+  - do:
+      get:
+        index: destination-index
+        id: doc_1
+
+  - match: { _source.sparse_field: "inference test" }
+  - match: { _source.dense_field: "another inference test" }
+  - match: { _source.non_inference_field: "non inference test" }
+
+  - do:
+      search:
+        index: destination-index
+        body:
+          _source:
+            exclude_vectors: false
+          query:
+            match_all: {}
+
+  - match: { hits.total.value: 1 }
+  - match: { hits.total.relation: eq }
+
+  - length: { hits.hits.0._source._inference_fields.sparse_field.inference.chunks: 1 }
+  - length: { hits.hits.0._source._inference_fields.sparse_field.inference.chunks.sparse_field: 1 }
+  - exists: hits.hits.0._source._inference_fields.sparse_field.inference.chunks.sparse_field.0.embeddings
+  - match: { hits.hits.0._source._inference_fields.sparse_field.inference.chunks.sparse_field.0.start_offset: 0 }
+  - match: { hits.hits.0._source._inference_fields.sparse_field.inference.chunks.sparse_field.0.end_offset: 14 }
+
+  - length: { hits.hits.0._source._inference_fields.dense_field.inference.chunks: 1 }
+  - length: { hits.hits.0._source._inference_fields.dense_field.inference.chunks.dense_field: 1 }
+  - exists: hits.hits.0._source._inference_fields.dense_field.inference.chunks.dense_field.0.embeddings
+  - match: { hits.hits.0._source._inference_fields.dense_field.inference.chunks.dense_field.0.start_offset: 0 }
+  - match: { hits.hits.0._source._inference_fields.dense_field.inference.chunks.dense_field.0.end_offset: 22 }
+
+---
+"Reindex recomputes embeddings by default":
+  - do:
+      index:
+        index: test-index
+        id: doc_1
+        body:
+          sparse_field: "inference test"
+          dense_field: "another inference test"
+          non_inference_field: "non inference test"
+        refresh: true
+
+  - do:
+      indices.create:
+        index: destination-index
+        body:
+          settings:
+            index:
+              mapping:
+                semantic_text:
+                  use_legacy_format: false
+          mappings:
+            properties:
+              sparse_field:
+                type: semantic_text
+                inference_id: dense-inference-id
+              dense_field:
+                type: semantic_text
+                inference_id: sparse-inference-id
+              non_inference_field:
+                type: text
+
+  # reindex fails when preserving embeddings
+  - do:
+      catch: bad_request
+      reindex:
+        wait_for_completion: true
+        body:
+          source:
+            index: test-index
+            _source:
+              exclude_vectors: false
+          dest:
+            index: destination-index
+
+  - length: { failures: 1 }
+
+  # reindex succeeds when embeddings are recomputed
+  - do:
+      reindex:
+        wait_for_completion: true
+        body:
+          source:
+            index: test-index
+          dest:
+            index: destination-index
+
+  - match: { created: 1 }
   - match: { updated: 0 }
   - match: { version_conflicts: 0 }
   - match: { failures: [ ] }

Original file line number	Diff line number	Diff line change
`@@ -214,7 +214,7 @@ static <Request extends AbstractBulkByScrollRequest<Request>> SearchRequest prep`
`214`	`214`	`// always include vectors in the response unless explicitly set`
`215`	`215`	`var fetchSource = sourceBuilder.fetchSource();`
`216`	`216`	`if (fetchSource == null) {`
`217`		`- sourceBuilder.fetchSource(FetchSourceContext.of(true, false, new String[0], new String[0]));`
	`217`	`+ sourceBuilder.fetchSource(FetchSourceContext.FETCH_ALL_SOURCE_EXCLUDE_INFERENCE_FIELDS);`
`218`	`218`	`} else if (fetchSource.excludeVectors() == null) {`
`219`	`219`	`sourceBuilder.excludeVectors(false);`
`220`	`220`	`}`