SparkRunner should deduplicate when an UnboundedSource requires Deduping

The implementation of an Unbounded Read does not inspect the requiresDeduping property of the source, and as such does not appropriately deduplicate sources that require it.

https://github.com/apache/beam/blob/master/runners/spark/src/main/java/org/apache/beam/runners/spark/io/SparkUnboundedSource.java

Imported from Jira [BEAM-1725](https://issues.apache.org/jira/browse/BEAM-1725). Original Jira may contain additional context.
Reported by: tgroh.

Start in runners/spark/src/main/java/org/apache/beam/runners/spark/io/SparkUnboundedSource.java and trace how an unbounded read handles a source's requiresDeduping property. Confirm the expected behavior for sources requiring deduplication and verify that the Spark runner deduplicates them without changing sources that do not require it.

Written by the indexing model from the issue text.