Python Language

Python Language

Priority: P0 (CRITICAL)

Rules

• Prefer from __future__ import annotations in modern modules.
• Type public functions, return values, and boundary data shapes.
• Use pathlib.Path, contextlib, enum, and stdlib helpers before custom wrappers.
• Use @dataclass(frozen=True) for immutable contracts and value objects.
• Prefer Protocol, TypedDict, or small dataclasses over dict[str, Any] sprawl.

Recipe

1. Model contracts explicitly with dataclasses, typed dicts, or pydantic at I/O boundaries.
2. Keep imports stable: stdlib, third-party, local.
3. Use comprehensions and generators carefully; stop when readability drops.
4. Choose explicit timezone-aware datetimes for persisted or user-visible timestamps.
5. Keep module side effects minimal; import should not start work.

Anti-Patterns

• No mutable defaults: use field(default_factory=...).
• No os.path in new code: prefer Path.
• No Any as escape hatch: narrow unknown shapes or validate them.
• No import-time env reads for volatile runtime state unless cached config is intentional.

Verify

• Run parser/type checks after cross-module type changes.
• Load Framework Map for category defaults.

References

• Framework Map