Skip to content

README.md

Latest commit

 

History

History
71 lines (51 loc) · 2.84 KB

README.md

File metadata and controls

71 lines (51 loc) · 2.84 KB
Raw

Chapter 5 — Data Class Builders

Chapter goal: Know exactly which data class builder to reach for and why. Understand the internals well enough to never be surprised by __eq__, __hash__, ClassVar, or field() behavior.

What You'll Learn

Section Core concept
Boilerplate problem Why plain classes break silently without __eq__/__repr__
namedtuple Factory, _fields, _asdict, _make, _replace, defaults
typing.NamedTuple Class syntax, typed fields, methods, runtime vs static type checking
@dataclass Decorator options: frozen, order, eq, unsafe_hash
field() default_factory, repr, compare, hash, init, metadata
__post_init__ Validation, computed fields, cross-field constraints
ClassVar / InitVar Class attributes vs init-only arguments

Key Files

  • examples.py — 7-part annotated walkthrough of all builders
  • exercises.py — 5 exercises: builder comparison, validated sensor, Money value object, JSON round-trip, typed DB rows
  • benchmarks.py — instantiation, memory, equality, hashing compared across all builders
  • mini_project.py — schema-validated config system using all three builders together
  • notes.md — deep technical reference with comparison tables and decision flowchart
  • pitfalls.md — 7 common bugs
  • interview_questions.md — 9 questions from L3→L6
  • architecture_notes.md — internals: how each builder generates methods

The Core Mental Model

Three builders, one goal: eliminate record class boilerplate.

  namedtuple:    tuple subclass → immutable, hashable, zero overhead
  NamedTuple:    namedtuple + type hints + class syntax + methods
  @dataclass:    regular class + generated methods + full flexibility

  All three auto-generate: __init__, __repr__, __eq__
  None enforce type hints at runtime (only static type checkers do)

Builder Selection — 30-Second Guide

Need mutable fields?          → @dataclass
Need dict/set key (hashable)? → namedtuple / NamedTuple / @dataclass(frozen=True)
Storing millions of records?  → namedtuple / NamedTuple (no __dict__ overhead)
Need validation in __init__?  → @dataclass with __post_init__
Value object (Money, Color)?  → NamedTuple (immutable, equality by value)
Quick throwaway record?       → namedtuple

Pre-Reading Checklist

Before starting, confirm you understand:

  • Why PlainCity() == PlainCity() is False without __eq__
  • Why class Broken: items: list = [] is a bug (shared mutable default)
  • What "tuple subclass" means for memory and API compatibility

Reference: Fluent Python 2nd ed., Chapter 5 — pages 163–216