python-expert-best-practices-code-review

Apply a prioritized rule set covering the most damaging patterns first: mutable default arguments, masked required dictionary keys, and generic except clauses. Each rule includes incorrect and correct code examples so reviewers can identify violations and suggest concrete fixes.

Use d[key] when the key is required so missing keys fail fast with KeyError at the source. Using dict.get() masks missing keys as None, causing errors to appear far from the root cause. Reserve dict.get(key, default) for genuinely optional keys with meaningful defaults.

Mutable defaults like [] or {} are evaluated once at function definition time and shared across all calls, so mutations persist between invocations. Use None as the default and create a new mutable object inside the function body instead.

Return NotImplemented in __add__ and similar dunder methods when the operand type is unsupported, allowing Python to try reverse operations like __radd__. In-place operators like __iadd__ must always return self and can accept broader types than their binary counterparts.

Opinionated rules cover common high-impact patterns but cannot catch domain-specific logic errors or architectural issues. They work best as a review checklist alongside testing and type checking, not as a replacement for them.