FP Python Standards (Module 09 – Team Adoption)

This repo’s defaults are functional by default:

• Pure core by default; effects only in funcpipe_rag.boundaries and funcpipe_rag.infra.
• Prefer stdlib FP (itertools, functools, operator, pathlib) before adding new abstractions.
• Keep configuration as immutable data; build pipelines from config (no global state).
• Keep error handling explicit via funcpipe_rag.result.types.Result and ErrInfo.
• Prefer small, composable functions; name intermediates when a pipeline becomes hard to scan.

Patterns

• Iterator pipelines: use itertools.chain, itertools.accumulate, itertools.islice, itertools.groupby (sorted/consecutive precondition).
• Higher-order: use functools.partial for configurators; use functools.lru_cache only for pure functions.
• Lambda avoidance: prefer operator.itemgetter/attrgetter for simple projections.
• Boundaries: shells/adapters interpret IOPlan/AsyncPlan; ports return descriptions, never execute.

Pragmatic escapes

• Loops are allowed where they improve clarity or performance.
• Eager materialization is allowed at explicit sinks/boundaries.
• Custom abstractions (monads/effects) exist from earlier modules; for Module 09 code, treat them as optional and prefer stdlib-first refactors unless there is a concrete benefit.