Apache Beam IO Connector

Hi folks 👋

I noted that there's an Apache Spark integration[1] for creating segments and batch-ingesting data. I'd be keen to explore creating a similar IO connector for Apache Beam[2], and I would be interested in providing support both in terms of brainstorming and/or implementation.

I'm opening this issue to better understand the community support of and implications of such an integration, what related work might already be underway, and how generalizable the core components of `org.apache.pinot.plugin.ingestion.batch.spark` might be.

[1] https://docs.pinot.apache.org/basics/data-import/batch-ingestion/spark
[2] https://beam.apache.org/documentation/io/developing-io-overview/

Review the Apache Spark batch-ingestion documentation and the org.apache.pinot.plugin.ingestion.batch.spark package, then compare them with Apache Beam's IO connector guidance. Clarify whether a Beam connector is already underway, which Spark components are generalizable, and what scope and acceptance criteria an implementation would need.

Written by the indexing model from the issue text.