Replace anonymous classes for SortOrder and FIlterExec overrides

[gerashegalov]

Demo PR contributing to #10838

It showcases a coding convention to follow using SortOrder and FilterExec replacements as an example

scala>  spark.range(100).where($"id" <= 10).collect()

java.lang.RuntimeException: convertToGpu failed
  at scala.sys.package$.error(package.scala:30)
  at com.nvidia.spark.rapids.GpuFilterExecMeta.convertToGpu(basicPhysicalOperators.scala:790)
  at com.nvidia.spark.rapids.GpuFilterExecMeta.convertToGpu(basicPhysicalOperators.scala:783)
  at com.nvidia.spark.rapids.SparkPlanMeta.convertIfNeeded(RapidsMeta.scala:838)
  at com.nvidia.spark.rapids.GpuOverrides$.com$nvidia$spark$rapids$GpuOverrides$$doConvertPlan(GpuOverrides.scala:4383)
  at com.nvidia.spark.rapids.GpuOverrides.applyOverrides(GpuOverrides.scala:4728)
  at com.nvidia.spark.rapids.GpuOverrides.$anonfun$applyWithContext$3(GpuOverrides.scala:4588)
  at com.nvidia.spark.rapids.GpuOverrides$.logDuration(GpuOverrides.scala:455)
  at com.nvidia.spark.rapids.GpuOverrides.$anonfun$applyWithContext$1(GpuOverrides.scala:4585)
  at com.nvidia.spark.rapids.GpuOverrideUtil$.$anonfun$tryOverride$1(GpuOverrides.scala:4551)
  at com.nvidia.spark.rapids.GpuOverrides.applyWithContext(GpuOverrides.scala:4605)
  at com.nvidia.spark.rapids.GpuOverrides.apply(GpuOverrides.scala:4578)
  at com.nvidia.spark.rapids.GpuOverrides.apply(GpuOverrides.scala:4574)
  at org.apache.spark.sql.execution.ApplyColumnarRulesAndInsertTransitions.$anonfun$apply$1(Columnar.scala:532)

Signed-off-by: Gera Shegalov [email protected]

[gerashegalov]

build

[jlowe]

I like the anonymous class changes, since we're trading class-for-class, making the stacktrace much more useful without adding to the classloading burden. As for the pluginChecks change, that I'm less convinced is a net benefit change.

[gerashegalov]

build