chapter_22_dynamic_attributes

Chapter 22 — Dynamic Attributes and Properties

Theme: Python allows you to dynamically intercept attribute lookups. This enables the creation of elegant, declarative APIs (like Django ORM or Pydantic) by transforming dictionary keys into dot-notation object attributes.

What You'll Learn

Concept	Explanation
__getattr__	Intercepts attribute lookups ONLY when the attribute does not exist.
__getattribute__	Intercepts EVERY attribute lookup unconditionally. Highly dangerous.
@property	Allows you to execute a function when a user accesses a basic attribute, enforcing the Uniform Access Principle.
__new__	The true object constructor. Executes before __init__ and physically returns the instance.

Key Files

• examples.py — Building a recursive JSON wrapper (FrozenJSON) from scratch using __getattr__.
• exercises.py — The infamous __getattr__ infinite recursion loop, and why you must sanitize Python keywords (like class) in APIs.
• mini_project.py — A robust API client wrapper that overrides __new__ to dynamically return either a list or a parsed object depending on the raw JSON type.
• benchmarks.py — Proves that metaprogramming adds a 2x–3x CPU overhead compared to accessing raw dictionary keys.
• notes.md — The mechanical difference between __getattr__ and __getattribute__.
• pitfalls.md — The severe architectural mistake of hiding heavy I/O operations (like database queries) behind simple @property lookups.
• interview_questions.md — L3 to L6 interview prep.
• architecture_notes.md — How functools.cached_property hacks the CPython __dict__ sequence to achieve O(1) performance.

30-Second Rules

# Rule 1: Always use __getattr__, NEVER use __getattribute__ unless you are an expert.
# __getattribute__ requires `super().__getattribute__(name)` to avoid infinite loops.

# Rule 2: If a JSON response contains a Python keyword, append an underscore.
if keyword.iskeyword(key):
    key += '_'  # 'class' becomes 'class_'

# Rule 3: Use @functools.cached_property for expensive dynamically calculated attributes.
@cached_property
def query_results(self):
    return run_large_db_query() # Only runs on the very first access!

Reference: Fluent Python 2nd ed., Chapter 22 — pages 863–905