Handle invalid cluster recommendation for Dataproc

[parthosa]

Fixes #1521.

Currently, AutoTuner/Bootstrapper recommends 1 x n1-standard-16 instance for the input CPU job, which used 8 cores and 2 instances. However, Dataproc does not support clusters with only one worker node.

This PR introduces validateRecommendedCluster, a validation mechanism for recommended cluster configurations. Platform-specific classes can override this method to enforce platform-specific constraints.

Changes

Enhancements to cluster recommendation validation:

• core/src/main/scala/com/nvidia/spark/rapids/tool/Platform.scala: Introduced the validateRecommendedCluster method to validate the recommended cluster configuration, allowing subclasses to provide platform-specific validation.
• core/src/main/scala/com/nvidia/spark/rapids/tool/Platform.scala: Implemented platform-specific validation in DataprocPlatform to ensure the number of worker nodes meets the minimum required by the platform.

Improvements to test coverage:

• core/src/test/scala/com/nvidia/spark/rapids/tool/qualification/QualificationSuite.scala: Modified tests to compare actual cluster information against expected values and added a new test to validate the recommended cluster information for invalid configurations. [1] [2] [3]
• core/src/test/scala/com/nvidia/spark/rapids/tool/qualification/QualificationSuite.scala: Refactored the runQualificationAndTestClusterInfo method to return the cluster summary, improving test readability and maintainability.

Test

• Added an invalid event log and a unit test for dataproc to verify that no cluster is recommended.
• In future, we could add similar constraints and unit tests for other platforms.