Python: Docs: Learn-site coverage for functional workflow API (PR #4238)

[moonbox3]

Summary

Track docs updates for the functional workflow API shipped in #4238 (closed #3537). Code, tests, and samples have landed; user-facing docs on the Learn site still need coverage so users can discover and adopt the new API alongside the existing graph-based WorkflowBuilder.

What shipped in #4238

• @workflow decorator — write workflows as plain async functions using native Python control flow (if/else, loops, asyncio.gather) instead of executor classes, edges, and builder patterns
• @step decorator — opt-in, saves results for HITL resume and crash recovery; plain functions work inside @workflow without it and simply re-execute on resume
• RunContext — optional parameter, needed only for HITL (ctx.request_info()), custom events (ctx.add_event()), or state (ctx.get_state() / ctx.set_state())
• Streaming — workflow.run(stream=True)
• HITL — ctx.request_info() with replay semantics
• .as_agent() — wraps a functional workflow as an agent-compatible object
• Samples under python/samples/01-get-started/ (05_functional_workflow_with_agents.py, 06_functional_workflow_basics.py) and python/samples/03-workflows/functional/ (basic_pipeline, basic_streaming_pipeline, parallel_pipeline, hitl_review, steps_and_checkpointing, agent_integration, naive_group_chat)

Docs work to do

• Add a Learn-site page introducing the functional workflow API as a stepping stone between single-agent use and the graph API
• Position functional vs. graph (WorkflowBuilder) API — when to pick which; both are supported
• Document @workflow usage, signatures with and without RunContext, and return/output semantics
• Document @step — what it does (result caching for HITL resume / crash recovery), when to use it (expensive operations: agent calls, API requests), and what happens without it (re-execution on resume)
• Document streaming (run(stream=True)) and event flow
• Document HITL with ctx.request_info() including replay behavior
• Document RunContext APIs: request_info, add_event, get_state / set_state
• Document .as_agent() and its use cases
• Link the new samples from the relevant Learn-site sections
• Cross-link from the existing workflow docs so users land on whichever API fits their task

References

• PR: Python: (core): Add functional workflow API #4238
• Parent issue: Python: Add functional workflow API #3537
• In-repo sample READMEs already updated: python/samples/01-get-started/README.md, python/samples/03-workflows/README.md, python/samples/README.md