[ES\QL] Text embedding function constant folding (#135710)

Author: afoucret

docs/reference/query-languages/esql/_snippets/functions/description/text_embedding.md

@@ -77,6 +77,7 @@
 import static org.elasticsearch.xpack.esql.action.EsqlCapabilities.Cap.RERANK;
 import static org.elasticsearch.xpack.esql.action.EsqlCapabilities.Cap.SEMANTIC_TEXT_FIELD_CAPS;
 import static org.elasticsearch.xpack.esql.action.EsqlCapabilities.Cap.SOURCE_FIELD_MAPPING;
+import static org.elasticsearch.xpack.esql.action.EsqlCapabilities.Cap.TEXT_EMBEDDING_FUNCTION;
 import static org.elasticsearch.xpack.esql.qa.rest.RestEsqlTestCase.assertNotPartial;
 import static org.elasticsearch.xpack.esql.qa.rest.RestEsqlTestCase.hasCapabilities;
 
@@ -224,7 +225,8 @@ protected boolean requiresInferenceEndpoint() {
             SEMANTIC_TEXT_FIELD_CAPS.capabilityName(),
             RERANK.capabilityName(),
             COMPLETION.capabilityName(),
-            KNN_FUNCTION_V5.capabilityName()
+            KNN_FUNCTION_V5.capabilityName(),
+            TEXT_EMBEDDING_FUNCTION.capabilityName()
         ).anyMatch(testCase.requiredCapabilities::contains);
     }
 

docs/reference/query-languages/esql/_snippets/functions/examples/text_embedding.md

@@ -1,15 +1,91 @@
-placeholder 
+text_embedding using a row source operator
 required_capability: text_embedding_function
-required_capability: not_existing_capability
+required_capability: dense_vector_field_type_released
 
 // tag::embedding-eval[]
 ROW input="Who is Victor Hugo?"
 | EVAL embedding = TEXT_EMBEDDING("Who is Victor Hugo?", "test_dense_inference")
 ;
 // end::embedding-eval[]
 
+input:keyword       | embedding:dense_vector
+Who is Victor Hugo? | [56.0, 50.0, 48.0]
+;
+
+
+text_embedding using a row source operator with query build using CONCAT
+required_capability: text_embedding_function
+required_capability: dense_vector_field_type_released
+
+ROW input="Who is Victor Hugo?"
+| EVAL embedding = TEXT_EMBEDDING(CONCAT("Who is ", "Victor Hugo?"), "test_dense_inference")
+;
 
 input:keyword       | embedding:dense_vector
 Who is Victor Hugo? | [56.0, 50.0, 48.0]
 ;
 
+
+text_embedding with knn on semantic_text_dense_field
+required_capability: text_embedding_function
+required_capability: dense_vector_field_type_released
+required_capability: knn_function_v5
+required_capability: semantic_text_field_caps
+
+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)
+| SORT _score DESC
+| LIMIT 10
+| KEEP semantic_text_field, query_embedding
+;
+
+semantic_text_field:text                                              | query_embedding:dense_vector
+be excellent to each other                                            | [45.0, 55.0, 54.0]          
+live long and prosper                                                 | [45.0, 55.0, 54.0]          
+all we have to decide is what to do with the time that is given to us | [45.0, 55.0, 54.0]          
+;
+
+text_embedding with knn (inline) on semantic_text_dense_field
+required_capability: text_embedding_function
+required_capability: dense_vector_field_type_released
+required_capability: knn_function_v5
+required_capability: semantic_text_field_caps
+
+FROM semantic_text METADATA _score
+| WHERE KNN(semantic_text_dense_field, TEXT_EMBEDDING("be excellent to each other", "test_dense_inference"))
+| SORT _score DESC
+| LIMIT 10
+| KEEP semantic_text_field
+;
+
+semantic_text_field:text                                             
+be excellent to each other                                           
+live long and prosper                                                
+all we have to decide is what to do with the time that is given to us
+;
+
+
+text_embedding with multiple knn queries in fork
+required_capability: text_embedding_function
+required_capability: dense_vector_field_type_released
+required_capability: knn_function_v5
+required_capability: fork_v9
+required_capability: semantic_text_field_caps
+
+FROM semantic_text METADATA _score
+| FORK (EVAL query_embedding = TEXT_EMBEDDING("be excellent to each other", "test_dense_inference") | WHERE KNN(semantic_text_dense_field, query_embedding))
+       (EVAL query_embedding = TEXT_EMBEDDING("live long and prosper", "test_dense_inference") | WHERE KNN(semantic_text_dense_field, query_embedding))
+| SORT _score DESC, _fork ASC
+| LIMIT 10
+| KEEP semantic_text_field, query_embedding, _fork
+;
+
+semantic_text_field:text                                              | query_embedding:dense_vector | _fork:keyword
+be excellent to each other                                            | [45.0, 55.0, 54.0]           | fork1        
+live long and prosper                                                 | [50.0, 57.0, 56.0]           | fork2        
+live long and prosper                                                 | [45.0, 55.0, 54.0]           | fork1        
+be excellent to each other                                            | [50.0, 57.0, 56.0]           | fork2        
+all we have to decide is what to do with the time that is given to us | [45.0, 55.0, 54.0]           | fork1        
+all we have to decide is what to do with the time that is given to us | [50.0, 57.0, 56.0]           | fork2        
+;

docs/reference/query-languages/esql/_snippets/functions/layout/text_embedding.md

@@ -88,7 +88,7 @@ public void esql(
             indexResolver,
             enrichPolicyResolver,
             preAnalyzer,
-            new LogicalPlanPreOptimizer(new LogicalPreOptimizerContext(foldContext)),
+            new LogicalPlanPreOptimizer(new LogicalPreOptimizerContext(foldContext, services.inferenceService())),
             functionRegistry,
             new LogicalPlanOptimizer(new LogicalOptimizerContext(cfg, foldContext)),
             mapper,