[ES|QL] Enable the TEXT_EMBEDDING function in non-snapshot build

docs/changelog/136103.yaml

@@ -0,0 +1,5 @@
+pr: 136103
+summary: Enable the TEXT_EMBEDDING function in non-snapshot build
+area: ES|QL
+type: feature
+issues: []

docs/reference/query-languages/esql/_snippets/lists/dense-vector-functions.md

@@ -1,4 +1,5 @@
 * {applies_to}`stack: preview 9.2` {applies_to}`serverless: preview` [`KNN`](../../functions-operators/dense-vector-functions.md#esql-knn)
+* {applies_to}`stack: preview 9.3` {applies_to}`serverless: preview` [`TEXT_EMBEDDING`](../../functions-operators/dense-vector-functions.md#esql-text_embedding)
 
 % * {applies_to}`stack: preview 9.3` {applies_to}`serverless: preview` [`V_COSINE`](../../functions-operators/dense-vector-functions.md#esql-v_cosine)
 % * {applies_to}`stack: preview 9.3` {applies_to}`serverless: preview` [`V_DOT_PRODUCT`](../../functions-operators/dense-vector-functions.md#esql-v_dot_product)

docs/reference/query-languages/esql/functions-operators/dense-vector-functions.md

@@ -23,6 +23,9 @@ and require appropriate field mappings.
 :::{include} ../_snippets/functions/layout/knn.md
 :::
 
+:::{include} ../_snippets/functions/layout/text_embedding.md
+:::
+
 % V_COSINE is currently a hidden feature
 % To make it visible again, uncomment this and the line in
 % lists/dense-vector-functions.md

x-pack/plugin/esql/qa/testFixtures/src/main/resources/text-embedding.csv-spec

@@ -2,11 +2,11 @@ text_embedding using a row source operator
 required_capability: text_embedding_function
 required_capability: dense_vector_field_type_released
 
-// tag::embedding-eval[]
+// tag::text-embedding-eval[]
 ROW input="Who is Victor Hugo?"
 | EVAL embedding = TEXT_EMBEDDING("Who is Victor Hugo?", "test_dense_inference")
+// end::text-embedding-eval[]
 ;
-// end::embedding-eval[]
 
 input:keyword       | embedding:dense_vector
 Who is Victor Hugo? | [56.0, 50.0, 48.0]
@@ -32,9 +32,11 @@ required_capability: dense_vector_field_type_released
 required_capability: knn_function_v5
 required_capability: semantic_text_field_caps
 
+// tag::text-embedding-knn[]
 FROM semantic_text METADATA _score
 | EVAL query_embedding = TEXT_EMBEDDING("be excellent to each other", "test_dense_inference")
 | WHERE KNN(semantic_text_dense_field, query_embedding)
+// end::text-embedding-knn[]
 | SORT _score DESC
 | LIMIT 10
 | KEEP semantic_text_field, query_embedding
@@ -52,8 +54,10 @@ required_capability: dense_vector_field_type_released
 required_capability: knn_function_v5
 required_capability: semantic_text_field_caps
 
+// tag::text-embedding-knn-inline[]
 FROM semantic_text METADATA _score
 | WHERE KNN(semantic_text_dense_field, TEXT_EMBEDDING("be excellent to each other", "test_dense_inference"))
+// end::text-embedding-knn-inline[]
 | SORT _score DESC
 | LIMIT 10
 | KEEP semantic_text_field

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/action/EsqlCapabilities.java

@@ -1243,7 +1243,7 @@ public enum Cap {
         /**
          * Support for the {@code TEXT_EMBEDDING} function for generating dense vector embeddings.
          */
-        TEXT_EMBEDDING_FUNCTION(Build.current().isSnapshot()),
+        TEXT_EMBEDDING_FUNCTION,
 
         /**
          * Support for the LIKE operator with a list of wildcards.

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/EsqlFunctionRegistry.java

@@ -526,7 +526,9 @@ private static FunctionDefinition[][] functions() {
                 def(AbsentOverTime.class, uni(AbsentOverTime::new), "absent_over_time"),
                 def(AvgOverTime.class, uni(AvgOverTime::new), "avg_over_time"),
                 def(LastOverTime.class, uni(LastOverTime::new), "last_over_time"),
-                def(FirstOverTime.class, uni(FirstOverTime::new), "first_over_time") } };
+                def(FirstOverTime.class, uni(FirstOverTime::new), "first_over_time"),
+                // dense vector function
+                def(TextEmbedding.class, bi(TextEmbedding::new), "text_embedding") } };
 
     }
 
@@ -545,8 +547,7 @@ private static FunctionDefinition[][] snapshotFunctions() {
                 def(L1Norm.class, L1Norm::new, "v_l1_norm"),
                 def(L2Norm.class, L2Norm::new, "v_l2_norm"),
                 def(Magnitude.class, Magnitude::new, "v_magnitude"),
-                def(Hamming.class, Hamming::new, "v_hamming"),
-                def(TextEmbedding.class, bi(TextEmbedding::new), "text_embedding") } };
+                def(Hamming.class, Hamming::new, "v_hamming") } };
     }
 
     public EsqlFunctionRegistry snapshotRegistry() {

x-pack/plugin/esql/src/main/java/org/elasticsearch/xpack/esql/expression/function/inference/TextEmbedding.java

@@ -40,23 +40,42 @@ public class TextEmbedding extends InferenceFunction<TextEmbedding> {
 
     @FunctionInfo(
         returnType = "dense_vector",
-        description = "Generates dense vector embeddings for text using a specified inference endpoint.",
-        appliesTo = { @FunctionAppliesTo(lifeCycle = FunctionAppliesToLifecycle.DEVELOPMENT) },
+        description = "Generates dense vector embeddings from text input using a specified "
+            + "[inference endpoint](docs-content://explore-analyze/elastic-inference/inference-api.md). "
+            + "Use this function to generate query vectors for KNN searches against your vectorized data "
+            + "or others dense vector based operations.",
+        appliesTo = { @FunctionAppliesTo(version = "9.3", lifeCycle = FunctionAppliesToLifecycle.PREVIEW) },
         preview = true,
         examples = {
             @Example(
-                description = "Generate text embeddings using the 'test_dense_inference' inference endpoint.",
+                description = "Basic text embedding generation from a text string using an inference endpoint.",
                 file = "text-embedding",
-                tag = "embedding-eval"
+                tag = "text-embedding-eval"
+            ),
+            @Example(
+                description = "Generate text embeddings and store them in a variable for reuse in KNN vector search queries.",
+                file = "text-embedding",
+                tag = "text-embedding-knn"
+            ),
+            @Example(
+                description = "Directly embed text within a KNN query for streamlined vector search without intermediate variables.",
+                file = "text-embedding",
+                tag = "text-embedding-knn-inline"
             ) }
     )
     public TextEmbedding(
         Source source,
-        @Param(name = "text", type = { "keyword" }, description = "Text to generate embeddings from") Expression inputText,
+        @Param(
+            name = "text",
+            type = { "keyword" },
+            description = "Text string to generate embeddings from. Must be a non-null literal string value."
+        ) Expression inputText,
         @Param(
             name = InferenceFunction.INFERENCE_ID_PARAMETER_NAME,
             type = { "keyword" },
-            description = "Identifier of the inference endpoint"
+            description = "Identifier of an existing inference endpoint the that will generate the embeddings. "
+                + "The inference endpoint must have the `text_embedding` task type and should use the same model "
+                + "that was used to embed your indexed data."
         ) Expression inferenceId
     ) {
         super(source, List.of(inputText, inferenceId));

x-pack/plugin/esql/src/test/java/org/elasticsearch/xpack/esql/analysis/AnalyzerTests.java

@@ -3814,8 +3814,6 @@ private void assertEmptyEsRelation(LogicalPlan plan) {
     }
 
     public void testTextEmbeddingResolveInferenceId() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         LogicalPlan plan = analyze(
             String.format(Locale.ROOT, """
                 FROM books METADATA _score | EVAL embedding = TEXT_EMBEDDING("italian food recipe", "%s")""", TEXT_EMBEDDING_INFERENCE_ID),
@@ -3833,8 +3831,6 @@ public void testTextEmbeddingResolveInferenceId() {
     }
 
     public void testTextEmbeddingFunctionResolveType() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         LogicalPlan plan = analyze(
             String.format(Locale.ROOT, """
                 FROM books METADATA _score| EVAL embedding = TEXT_EMBEDDING("italian food recipe", "%s")""", TEXT_EMBEDDING_INFERENCE_ID),
@@ -3853,8 +3849,6 @@ public void testTextEmbeddingFunctionResolveType() {
     }
 
     public void testTextEmbeddingFunctionMissingInferenceIdError() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         VerificationException ve = expectThrows(
             VerificationException.class,
             () -> analyze(
@@ -3868,8 +3862,6 @@ public void testTextEmbeddingFunctionMissingInferenceIdError() {
     }
 
     public void testTextEmbeddingFunctionInvalidInferenceIdError() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         String inferenceId = randomInferenceIdOtherThan(TEXT_EMBEDDING_INFERENCE_ID);
         VerificationException ve = expectThrows(
             VerificationException.class,
@@ -3887,8 +3879,6 @@ public void testTextEmbeddingFunctionInvalidInferenceIdError() {
     }
 
     public void testTextEmbeddingFunctionWithoutModel() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         ParsingException ve = expectThrows(ParsingException.class, () -> analyze("""
             FROM books METADATA _score| EVAL embedding = TEXT_EMBEDDING("italian food recipe")""", "mapping-books.json"));
 
@@ -3899,8 +3889,6 @@ public void testTextEmbeddingFunctionWithoutModel() {
     }
 
     public void testKnnFunctionWithTextEmbedding() {
-        assumeTrue("TEXT_EMBEDDING function required", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-
         LogicalPlan plan = analyze(
             String.format(Locale.ROOT, """
                 from test | where KNN(float_vector, TEXT_EMBEDDING("italian food recipe", "%s"))""", TEXT_EMBEDDING_INFERENCE_ID),

x-pack/plugin/esql/src/test/java/org/elasticsearch/xpack/esql/analysis/VerifierTests.java

@@ -2735,7 +2735,6 @@ public void testSortInTimeSeries() {
     }
 
     public void testTextEmbeddingFunctionInvalidQuery() {
-        assumeTrue("TEXT_EMBEDDING is not enabled", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
         assertThat(
             error("from test | EVAL embedding = TEXT_EMBEDDING(null, ?)", defaultAnalyzer, TEXT_EMBEDDING_INFERENCE_ID),
             equalTo("1:30: first argument of [TEXT_EMBEDDING(null, ?)] cannot be null, received [null]")
@@ -2753,7 +2752,6 @@ public void testTextEmbeddingFunctionInvalidQuery() {
     }
 
     public void testTextEmbeddingFunctionInvalidInferenceId() {
-        assumeTrue("TEXT_EMBEDDING is not enabled", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
         assertThat(
             error("from test | EVAL embedding = TEXT_EMBEDDING(?, null)", defaultAnalyzer, "query text"),
             equalTo("1:30: second argument of [TEXT_EMBEDDING(?, null)] cannot be null, received [null]")

x-pack/plugin/esql/src/test/java/org/elasticsearch/xpack/esql/expression/function/inference/TextEmbeddingErrorTests.java

@@ -7,7 +7,6 @@
 
 package org.elasticsearch.xpack.esql.expression.function.inference;
 
-import org.elasticsearch.xpack.esql.action.EsqlCapabilities;
 import org.elasticsearch.xpack.esql.core.expression.Expression;
 import org.elasticsearch.xpack.esql.core.expression.TypeResolutions;
 import org.elasticsearch.xpack.esql.core.tree.Source;
@@ -16,7 +15,6 @@
 import org.elasticsearch.xpack.esql.expression.function.ErrorsForCasesWithoutExamplesTestCase;
 import org.elasticsearch.xpack.esql.expression.function.TestCaseSupplier;
 import org.hamcrest.Matcher;
-import org.junit.Before;
 
 import java.util.List;
 import java.util.Locale;
@@ -28,12 +26,6 @@
  * Tests error conditions and type validation for TEXT_EMBEDDING function.
  */
 public class TextEmbeddingErrorTests extends ErrorsForCasesWithoutExamplesTestCase {
-
-    @Before
-    public void checkCapability() {
-        assumeTrue("TEXT_EMBEDDING is not enabled", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-    }
-
     @Override
     protected List<TestCaseSupplier> cases() {
         return paramsToSuppliers(TextEmbeddingTests.parameters());

x-pack/plugin/esql/src/test/java/org/elasticsearch/xpack/esql/expression/function/inference/TextEmbeddingTests.java

@@ -10,14 +10,12 @@
 import com.carrotsearch.randomizedtesting.annotations.Name;
 import com.carrotsearch.randomizedtesting.annotations.ParametersFactory;
 
-import org.elasticsearch.xpack.esql.action.EsqlCapabilities;
 import org.elasticsearch.xpack.esql.core.expression.Expression;
 import org.elasticsearch.xpack.esql.core.tree.Source;
 import org.elasticsearch.xpack.esql.expression.function.AbstractFunctionTestCase;
 import org.elasticsearch.xpack.esql.expression.function.FunctionName;
 import org.elasticsearch.xpack.esql.expression.function.TestCaseSupplier;
 import org.hamcrest.Matchers;
-import org.junit.Before;
 
 import java.util.List;
 import java.util.function.Supplier;
@@ -28,11 +26,6 @@
 
 @FunctionName("text_embedding")
 public class TextEmbeddingTests extends AbstractFunctionTestCase {
-    @Before
-    public void checkCapability() {
-        assumeTrue("TEXT_EMBEDDING is not enabled", EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION.isEnabled());
-    }
-
     public TextEmbeddingTests(@Name("TestCase") Supplier<TestCaseSupplier.TestCase> testCaseSupplier) {
         this.testCase = testCaseSupplier.get();
     }