Dear Yi Hu, I hope you are doing well.
I would like to kindly request your review of my pull request for the Apache Beam project: PR: https://github.com/apache/beam/pull/37747/ This contribution is part of my work related to the *Apache Beam Python SDK native streaming transforms* effort. In this PR, I introduced the foundational API for *Unbounded Sources* in the Python SDK, which aims to provide core abstractions for continuous data stream processing. The implementation currently focuses on the initial building blocks, including components such as CheckpointMark and UnboundedReader, which are designed to support checkpointing, fault tolerance, and continuous data ingestion. Further integration with Splittable DoFn (SDF) and full Read transform support will be implemented in subsequent steps. I would greatly appreciate it if you could review the changes and let me know if any improvements or modifications are needed. Your feedback will help ensure the design aligns well with the Python SDK architecture and the broader Beam ecosystem. I would also like to thank the Apache Beam community for their guidance and support throughout this contribution process. Looking forward to your feedback. Best regards, Shaheer Amjad
