Implement Spark DataSource Support for querying bootstrapped tables

We need ability in SparkDataSource to query COW table which is bootstrapped as per https://cwiki.apache.org/confluence/display/HUDI/RFC+-+12+:+Efficient+Migration+of+Large+Parquet+Tables+to+Apache+Hudi#RFC-12:EfficientMigrationofLargeParquetTablestoApacheHudi-BootstrapIndex:

Current implementation delegates to Parquet DataSource but this wont work as we need ability to stitch the columns externally.