I have an existing Apache Beam pipeline written in Python that consumes streaming data from Google Pub/Sub. I’m comfortable with the ingestion and sink stages, but I need clear, practical guidance on the transformation layer. What I need from you: • Review the current transformation PTransforms and explain—step-by-step—how the data is being mapped, filtered, windowed, or aggregated. • Suggest and implement simple code adjustments that achieve the tweaks I describe (for example, changing a windowing strategy or adding an extra enrichment step). • Deliver clean, well-commented code snippets so I can understand the logic and extend it myself later. I’m looking for a concise walkthrough rather than a full rewrite—think annotated code diff, quick pointers on best practices, and a brief Q&A session to make sure I can maintain the pipeline on my own. If you’re fluent in Python Beam and familiar with streaming patterns on Pub/Sub, let’s get started.