Merge pull request #29883 from vespa-engine/balder/drop-tokenize-for-streaming

Drop tokenize expressions from ilscript for streaming mode.

7 files changed, 72 insertions, 11 deletions

diff --git a/config-model/src/main/java/com/yahoo/schema/Schema.java b/config-model/src/main/java/com/yahoo/schema/Schema.java
index 279b5729ea1..3402ba31be9 100644
--- a/config-model/src/main/java/com/yahoo/schema/Schema.java
+++ b/config-model/src/main/java/com/yahoo/schema/Schema.java
@@ -18,7 +18,6 @@ import com.yahoo.schema.document.SDDocumentType;
 import com.yahoo.schema.document.SDField;
 import com.yahoo.schema.document.Stemming;
 import com.yahoo.schema.document.TemporaryImportedFields;
-import com.yahoo.schema.document.annotation.SDAnnotationType;
 import com.yahoo.searchlib.rankingexpression.Reference;
 import com.yahoo.vespa.documentmodel.DocumentSummary;
 import com.yahoo.vespa.documentmodel.SummaryField;
diff --git a/config-model/src/main/java/com/yahoo/schema/derived/DerivedConfiguration.java b/config-model/src/main/java/com/yahoo/schema/derived/DerivedConfiguration.java
index 7f874d04f13..575b7264628 100644
--- a/config-model/src/main/java/com/yahoo/schema/derived/DerivedConfiguration.java
+++ b/config-model/src/main/java/com/yahoo/schema/derived/DerivedConfiguration.java
@@ -80,7 +80,7 @@ public class DerivedConfiguration implements AttributesConfig.Producer {
                 summaries = new Summaries(schema, deployState.getDeployLogger(), deployState.getProperties().featureFlags());
                 juniperrc = new Juniperrc(schema);
                 rankProfileList = new RankProfileList(schema, schema.rankExpressionFiles(), attributeFields, deployState);
-                indexingScript = new IndexingScript(schema);
+                indexingScript = new IndexingScript(schema, isStreaming);
                 indexInfo = new IndexInfo(schema, isStreaming);
                 schemaInfo = new SchemaInfo(schema, deployState.rankProfileRegistry(), summaries);
                 indexSchema = new IndexSchema(schema);
diff --git a/config-model/src/main/java/com/yahoo/schema/derived/IndexingScript.java b/config-model/src/main/java/com/yahoo/schema/derived/IndexingScript.java
index 73f3507ab00..6f0420f5203 100644
--- a/config-model/src/main/java/com/yahoo/schema/derived/IndexingScript.java
+++ b/config-model/src/main/java/com/yahoo/schema/derived/IndexingScript.java
@@ -6,6 +6,7 @@ import com.yahoo.schema.document.GeoPos;
 import com.yahoo.schema.document.ImmutableSDField;
 import com.yahoo.vespa.configdefinition.IlscriptsConfig;
 import com.yahoo.vespa.configdefinition.IlscriptsConfig.Ilscript.Builder;
+import com.yahoo.vespa.indexinglanguage.ExpressionConverter;
 import com.yahoo.vespa.indexinglanguage.ExpressionVisitor;
 import com.yahoo.vespa.indexinglanguage.expressions.ClearStateExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.Expression;
@@ -16,6 +17,7 @@ import com.yahoo.vespa.indexinglanguage.expressions.PassthroughExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.ScriptExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.SetLanguageExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.StatementExpression;
+import com.yahoo.vespa.indexinglanguage.expressions.TokenizeExpression;
 import com.yahoo.vespa.indexinglanguage.expressions.ZCurveExpression;
 
 import java.util.ArrayList;
@@ -36,8 +38,10 @@ public final class IndexingScript extends Derived implements IlscriptsConfig.Pro
     private final List<String> docFields = new ArrayList<>();
     private final List<Expression> expressions = new ArrayList<>();
     private List<ImmutableSDField> fieldsSettingLanguage;
+    private final boolean isStreaming;
 
-    public IndexingScript(Schema schema) {
+    public IndexingScript(Schema schema, boolean isStreaming) {
+        this.isStreaming = isStreaming;
         derive(schema);
     }
 
@@ -96,21 +100,38 @@ public final class IndexingScript extends Derived implements IlscriptsConfig.Pro
         configBuilder.ilscript(ilscriptBuilder);
     }
 
+    private static class DropTokenize extends ExpressionConverter {
+        @Override
+        protected boolean shouldConvert(Expression exp) {
+            return exp instanceof TokenizeExpression;
+        }
+
+        @Override
+        protected Expression doConvert(Expression exp) {
+            return null;
+        }
+    }
+
     private void addContentInOrder(IlscriptsConfig.Ilscript.Builder ilscriptBuilder) {
         ArrayList<Expression> later = new ArrayList<>();
         Set<String> touchedFields = new HashSet<>();
         for (Expression expression : expressions) {
-            if (modifiesSelf(expression) && ! setsLanguage(expression))
+            if (isStreaming) {
+                expression = expression.convertChildren(new DropTokenize());
+            }
+            if (modifiesSelf(expression) && ! setsLanguage(expression)) {
                 later.add(expression);
-            else
+            } else {
                 ilscriptBuilder.content(expression.toString());
+            }
 
             FieldScanVisitor fieldFetcher = new FieldScanVisitor();
             fieldFetcher.visit(expression);
             touchedFields.addAll(fieldFetcher.touchedFields());
         }
-        for (Expression exp : later)
+        for (Expression exp : later) {
             ilscriptBuilder.content(exp.toString());
+        }
         generateSyntheticStatementsForUntouchedFields(ilscriptBuilder, touchedFields);
     }
 
@@ -171,8 +192,8 @@ public final class IndexingScript extends Derived implements IlscriptsConfig.Pro
     }
 
     private static class FieldScanVisitor extends ExpressionVisitor {
-        List<String> touchedFields = new ArrayList<String>();
-        List<String> candidates = new ArrayList<String>();
+        List<String> touchedFields = new ArrayList<>();
+        List<String> candidates = new ArrayList<>();
 
         @Override
         protected void doVisit(Expression exp) {
diff --git a/config-model/src/test/java/com/yahoo/schema/AttributeSettingsTestCase.java b/config-model/src/test/java/com/yahoo/schema/AttributeSettingsTestCase.java
index ceac2b94997..6624a7d3b32 100644
--- a/config-model/src/test/java/com/yahoo/schema/AttributeSettingsTestCase.java
+++ b/config-model/src/test/java/com/yahoo/schema/AttributeSettingsTestCase.java
@@ -280,7 +280,7 @@ public class AttributeSettingsTestCase extends AbstractSchemaTestCase {
 
     @Test
     void requireThatMutableIsAllowedThroughIndexing() throws ParseException {
-        IndexingScript script = new IndexingScript(getSearchWithMutables());
+        IndexingScript script = new IndexingScript(getSearchWithMutables(), false);
         IlscriptsConfig.Builder builder = new IlscriptsConfig.Builder();
         script.getConfig(builder);
         IlscriptsConfig cfg = builder.build();
diff --git a/config-model/src/test/java/com/yahoo/schema/derived/IndexingScriptTestCase.java b/config-model/src/test/java/com/yahoo/schema/derived/IndexingScriptTestCase.java
new file mode 100644
index 00000000000..81536bf99fb
--- /dev/null
+++ b/config-model/src/test/java/com/yahoo/schema/derived/IndexingScriptTestCase.java
@@ -0,0 +1,41 @@
+// Copyright Vespa.ai. Licensed under the terms of the Apache 2.0 license. See LICENSE in the project root.
+package com.yahoo.schema.derived;
+
+import com.yahoo.document.DataType;
+import com.yahoo.schema.Schema;
+import com.yahoo.schema.document.MatchType;
+import com.yahoo.schema.document.Matching;
+import com.yahoo.schema.document.SDField;
+import com.yahoo.schema.document.TemporarySDField;
+import com.yahoo.vespa.configdefinition.IlscriptsConfig;
+import org.junit.jupiter.api.Test;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+
+public class IndexingScriptTestCase {
+
+    private static IlscriptsConfig ilscriptsConfig(Schema schema, boolean isStreaming) {
+        IndexingScript script = new IndexingScript(schema, isStreaming);
+        IlscriptsConfig.Builder cfgBuilder = new IlscriptsConfig.Builder();
+        script.getConfig(cfgBuilder);
+        return cfgBuilder.build();
+    }
+
+    private void verifyIndexingScript(boolean isStreaming) {
+        Schema schema = VsmFieldsTestCase.createSchema();
+        SDField field = new TemporarySDField(schema.getDocument(), "f", DataType.STRING);
+        field.parseIndexingScript("{ tokenize | index }");
+        field.setMatching(new Matching(MatchType.TEXT));
+        schema.getDocument().addField(field);
+        IlscriptsConfig cfg = ilscriptsConfig(schema, isStreaming);
+        assertEquals(1, cfg.ilscript().size());
+        assertEquals(2, cfg.ilscript(0).content().size());
+        String indexing = isStreaming ? "index" : "tokenize | index";
+        assertEquals("clear_state | guard { " + indexing + "; }", cfg.ilscript(0).content(0));
+    }
+    @Test
+    void testThatTokenizeIsIgnoredFromStreaming() {
+        verifyIndexingScript(false);
+        verifyIndexingScript(true);
+    }
+}
diff --git a/config-model/src/test/java/com/yahoo/schema/derived/VsmFieldsTestCase.java b/config-model/src/test/java/com/yahoo/schema/derived/VsmFieldsTestCase.java
index e469b80ab40..042feeb5f46 100644
--- a/config-model/src/test/java/com/yahoo/schema/derived/VsmFieldsTestCase.java
+++ b/config-model/src/test/java/com/yahoo/schema/derived/VsmFieldsTestCase.java
@@ -23,7 +23,7 @@ import static org.junit.jupiter.api.Assertions.assertEquals;
  */
 public class VsmFieldsTestCase {
 
-    private static Schema createSchema() {
+     static Schema createSchema() {
         Schema schema = new Schema("test", MockApplicationPackage.createEmpty(), new MockFileRegistry(), new TestableDeployLogger(), new TestProperties());
         var sdoc = new SDDocumentType("test");
         schema.addDocument(sdoc);
diff --git a/config-model/src/test/java/com/yahoo/schema/processing/AssertIndexingScript.java b/config-model/src/test/java/com/yahoo/schema/processing/AssertIndexingScript.java
index 8aeb11aaa2e..b82dc197838 100644
--- a/config-model/src/test/java/com/yahoo/schema/processing/AssertIndexingScript.java
+++ b/config-model/src/test/java/com/yahoo/schema/processing/AssertIndexingScript.java
@@ -18,7 +18,7 @@ import static org.junit.jupiter.api.Assertions.fail;
 public abstract class AssertIndexingScript {
 
     public static void assertIndexing(List<String> expected, Schema schema) {
-        assertIndexing(expected, new IndexingScript(schema).expressions());
+        assertIndexing(expected, new IndexingScript(schema, false).expressions());
     }
 
     public static void assertIndexing(List<String> expected, IndexingScript script) {