Which issue does this PR close?

TBD

Rationale for this change

Flink's BATCH mode currently has no native execution: DefaultPlannerFactory routes it to the stock, unmodified BatchPlanner. NativePlanFusionBuilder and FlinkAuronCalcOperator are already engine-agnostic (no streaming-specific types), so the smallest correct first step is a BatchExecCalc shadow reusing both, mirroring the existing StreamExecCalc shadow.

What changes are included in this PR?

• New org.apache.flink.table.planner.plan.nodes.exec.batch.BatchExecCalc in auron-flink-planner, matching Flink 1.18.1's stock constructor signature exactly.
  • Native conversion via NativePlanFusionBuilder.buildNativeCalcPlan, falling back to Flink's codegen Calc (or throwing if fallback is disabled), same as streaming.
  • No graph-level source fusion yet (no batch counterpart to AuronOperatorFusionProcessor); supportFusionCodegen() returns false.
  • No @ExecNodeMetadata/@JsonCreator, matching stock batch BatchExecCalc at this Flink version (compiled-plan restoration is streaming-only here).
• Registered in shadowed-flink-execnodes.txt; added matching shade <exclude> in auron-flink-assembly/pom.xml.

No changes to any streaming file or to DefaultPlannerFactory.

Are there any user-facing changes?

Yes: BATCH queries now get native execution for Calc (projection/filter) when auron-flink-assembly is ahead of stock flink-table-planner on the classpath. Falls back to Flink's codegen Calc when not convertible, matching existing streaming behavior.

How was this patch tested?

• ShadowedExecNodeRegistryTest and AssemblyJarStructureIT pass with the new class registered and shaded correctly.
• Verified constructor signature and lack of @ExecNodeMetadata/@JsonCreator against the real flink-table-planner_2.12-1.18.1.jar.
• Manually verified end-to-end on a local Flink 1.18.1 cluster: a BATCH SQL query's Calc step executes via Auron's native engine (native ProjectExec/FFIReader in task logs).

Draft while a dedicated BatchExecCalcTest (mirroring StreamExecCalcTest) is added.

Was this patch authored or co-authored using generative AI tooling?

• Yes
• No

ASF guidance: https://www.apache.org/legal/generative-tooling.html