Add micro benchmarks for unified query layer

benchmarks/README.md

@@ -10,7 +10,18 @@ The microbenchmark suite is also handy for ad-hoc microbenchmarks but please rem
 
 ## Getting Started
 
-Just run `./gradlew :benchmarks:jmh` from the project root directory or run specific benchmark via your IDE. It will build all microbenchmarks, execute them and print the result.
+Run all benchmarks from the project root directory:
+
+```bash
+./gradlew :benchmarks:jmh
+```
+
+Run specific benchmarks using the `-Pjmh.includes` parameter:
+
+```bash
+./gradlew :benchmarks:jmh -Pjmh.includes='UnifiedQueryBenchmark'
+./gradlew :benchmarks:jmh -Pjmh.includes='UnifiedQueryBenchmark.plan.*'
+```
 
 ## Adding Microbenchmarks
 

benchmarks/build.gradle

@@ -5,6 +5,7 @@
 
 plugins {
     id 'java-library'
+    id "io.freefair.lombok"
     id "me.champeau.jmh" version "0.7.3"
 }
 
@@ -15,6 +16,8 @@ repositories {
 dependencies {
     implementation project(':core')
     implementation project(':opensearch')
+    implementation project(':api')
+    jmhImplementation testFixtures(project(':api'))
 
     // Dependencies required by JMH micro benchmark
     api group: 'org.openjdk.jmh', name: 'jmh-core', version: '1.36'
@@ -30,4 +33,9 @@ spotless {
     }
 }
 
+// JMH configuration passed via command line
+jmh {
+    includes = project.hasProperty('jmh.includes') ? [project.property('jmh.includes')] : []
+}
+
 compileJava.options.compilerArgs.addAll(["-processor", "org.openjdk.jmh.generators.BenchmarkProcessor"])

benchmarks/src/jmh/java/org/opensearch/sql/api/UnifiedFunctionBenchmark.java

@@ -0,0 +1,128 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.opensearch.sql.api;
+
+import java.util.Arrays;
+import java.util.List;
+import java.util.concurrent.TimeUnit;
+import lombok.Getter;
+import lombok.RequiredArgsConstructor;
+import org.openjdk.jmh.annotations.Benchmark;
+import org.openjdk.jmh.annotations.BenchmarkMode;
+import org.openjdk.jmh.annotations.Fork;
+import org.openjdk.jmh.annotations.Level;
+import org.openjdk.jmh.annotations.Measurement;
+import org.openjdk.jmh.annotations.Mode;
+import org.openjdk.jmh.annotations.OperationsPerInvocation;
+import org.openjdk.jmh.annotations.OutputTimeUnit;
+import org.openjdk.jmh.annotations.Param;
+import org.openjdk.jmh.annotations.Scope;
+import org.openjdk.jmh.annotations.Setup;
+import org.openjdk.jmh.annotations.State;
+import org.openjdk.jmh.annotations.TearDown;
+import org.openjdk.jmh.annotations.Warmup;
+import org.openjdk.jmh.infra.Blackhole;
+import org.opensearch.sql.api.function.UnifiedFunction;
+import org.opensearch.sql.api.function.UnifiedFunctionRepository;
+
+/**
+ * JMH benchmark for measuring {@link UnifiedFunction} performance. Tests one representative
+ * function per PPL category (json, math, conditional, collection, string).
+ *
+ * <p>Benchmarks:
+ *
+ * <ul>
+ *   <li>{@link #loadFunction()}: Measures function loading from repository
+ *   <li>{@link #evalFunction(Blackhole)}: Measures function evaluation (1000 calls per invocation)
+ * </ul>
+ */
+@Warmup(iterations = 2, time = 1)
+@Measurement(iterations = 5, time = 1)
+@BenchmarkMode(Mode.AverageTime)
+@OutputTimeUnit(TimeUnit.MILLISECONDS)
+@State(Scope.Thread)
+@Fork(value = 1)
+public class UnifiedFunctionBenchmark extends UnifiedQueryTestBase {
+
+  /** Number of function evaluations per benchmark invocation. */
+  private static final int OPS = 1000;
+
+  /** Benchmark specification selecting which function to test. */
+  @Param public BenchmarkSpec benchmarkSpec;
+
+  /** Repository for loading unified functions. */
+  private UnifiedFunctionRepository repository;
+
+  /** Pre-loaded function instance for evaluation benchmarks. */
+  private UnifiedFunction function;
+
+  /** Input arguments for function evaluation. */
+  private List<Object> inputs;
+
+  @Setup(Level.Trial)
+  public void setUpBenchmark() {
+    super.setUp();
+
+    repository = new UnifiedFunctionRepository(context);
+    function = benchmarkSpec.loadFunction(repository);
+    inputs = benchmarkSpec.getInputs();
+  }
+
+  @TearDown(Level.Trial)
+  public void tearDownBenchmark() throws Exception {
+    super.tearDown();
+  }
+
+  /** Benchmarks function loading from repository. */
+  @Benchmark
+  public UnifiedFunction loadFunction() {
+    return benchmarkSpec.loadFunction(repository);
+  }
+
+  /** Benchmarks function evaluation with pre-loaded function and inputs. */
+  @Benchmark
+  @OperationsPerInvocation(OPS)
+  public void evalFunction(Blackhole bh) {
+    for (int i = 0; i < OPS; i++) {
+      bh.consume(function.eval(inputs));
+    }
+  }
+
+  /** Enum defining benchmark test cases - one representative function per PPL category. */
+  @RequiredArgsConstructor
+  public enum BenchmarkSpec {
+    JSON_EXTRACT(
+        inputTypes("VARCHAR", "VARCHAR"),
+        sampleInputs("{\"name\":\"test\",\"value\":42}", "$.name")),
+    COALESCE(
+        inputTypes("VARCHAR", "VARCHAR", "VARCHAR"),
+        sampleInputs(null, "first_value", "default_value")),
+    MVFIND(
+        inputTypes("ARRAY", "VARCHAR"),
+        sampleInputs(List.of("debug", "error", "warn", "info"), "err.*")),
+    REX_EXTRACT(
+        inputTypes("VARCHAR", "VARCHAR", "INTEGER"),
+        sampleInputs("192.168.1.1 - GET /api", "(\\d+)", 1));
+
+    private final List<String> inputTypes;
+    @Getter private final List<Object> inputs;
+
+    UnifiedFunction loadFunction(UnifiedFunctionRepository repository) {
+      return repository
+          .loadFunction(name())
+          .map(desc -> desc.getBuilder().build(inputTypes))
+          .orElseThrow();
+    }
+
+    private static List<String> inputTypes(String... types) {
+      return List.of(types);
+    }
+
+    private static List<Object> sampleInputs(Object... args) {
+      return Arrays.asList(args);
+    }
+  }
+}

benchmarks/src/jmh/java/org/opensearch/sql/api/UnifiedQueryBenchmark.java

@@ -0,0 +1,87 @@
+/*
+ * Copyright OpenSearch Contributors
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+package org.opensearch.sql.api;
+
+import java.sql.PreparedStatement;
+import java.util.concurrent.TimeUnit;
+import org.apache.calcite.rel.RelNode;
+import org.apache.calcite.sql.dialect.SparkSqlDialect;
+import org.openjdk.jmh.annotations.Benchmark;
+import org.openjdk.jmh.annotations.BenchmarkMode;
+import org.openjdk.jmh.annotations.Fork;
+import org.openjdk.jmh.annotations.Level;
+import org.openjdk.jmh.annotations.Measurement;
+import org.openjdk.jmh.annotations.Mode;
+import org.openjdk.jmh.annotations.OutputTimeUnit;
+import org.openjdk.jmh.annotations.Param;
+import org.openjdk.jmh.annotations.Scope;
+import org.openjdk.jmh.annotations.Setup;
+import org.openjdk.jmh.annotations.State;
+import org.openjdk.jmh.annotations.TearDown;
+import org.openjdk.jmh.annotations.Warmup;
+import org.opensearch.sql.api.compiler.UnifiedQueryCompiler;
+import org.opensearch.sql.api.transpiler.UnifiedQueryTranspiler;
+
+/**
+ * JMH benchmark for measuring the overhead of unified query API components when processing queries.
+ * This provides baseline metrics and guidance for API consumers during integration.
+ */
+@Warmup(iterations = 2, time = 1)
+@Measurement(iterations = 5, time = 1)
+@BenchmarkMode(Mode.AverageTime)
+@OutputTimeUnit(TimeUnit.MILLISECONDS)
+@State(Scope.Thread)
+@Fork(value = 1)
+public class UnifiedQueryBenchmark extends UnifiedQueryTestBase {
+
+  /** Common query patterns for benchmarking. */
+  @Param({
+    "source = catalog.employees",
+    "source = catalog.employees | where age > 30",
+    "source = catalog.employees | stats count() by department",
+    "source = catalog.employees | sort - age",
+    "source = catalog.employees | where age > 25 | stats avg(age) by department | sort - department"
+  })
+  private String query;
+
+  /** Transpiler for converting logical plans to SQL strings. */
+  private UnifiedQueryTranspiler transpiler;
+
+  /** Compiler for converting logical plans to executable statements. */
+  private UnifiedQueryCompiler compiler;
+
+  @Setup(Level.Trial)
+  public void setUpBenchmark() {
+    super.setUp();
+    transpiler = UnifiedQueryTranspiler.builder().dialect(SparkSqlDialect.DEFAULT).build();
+    compiler = new UnifiedQueryCompiler(context);
+  }
+
+  @TearDown(Level.Trial)
+  public void tearDownBenchmark() throws Exception {
+    super.tearDown();
+  }
+
+  /** Benchmarks query parsing and Calcite logical plan generation. */
+  @Benchmark
+  public RelNode planQuery() {
+    return planner.plan(query);
+  }
+
+  /** Benchmarks the full transpilation pipeline: Query → logical plan → SQL string. */
+  @Benchmark
+  public String transpileToSql() {
+    RelNode plan = planner.plan(query);
+    return transpiler.toSql(plan);
+  }
+
+  /** Benchmarks the compilation pipeline: Query → logical plan → executable statement. */
+  @Benchmark
+  public PreparedStatement compileQuery() {
+    RelNode plan = planner.plan(query);
+    return compiler.compile(plan);
+  }
+}