python-try-except

Grep for `try:` across your files, then rank each except clause from most to least dangerous: bare `except:`, then `except Exception:`, then tuples, then specific types. Each caught type should have a clear justification tied to what the try block can actually raise.

Use conditional checks for local values you control: `if key in d` or `d.get(key)` instead of catching KeyError, `hasattr(x, "y")` instead of catching AttributeError, and length checks instead of IndexError. Reserve try/except for external state like filesystem, network, and parsing operations.

Move setup statements above the try and processing logic after it or into an `else` clause, so the try contains only the single operation that can raise the expected exception. Nested try/except blocks often signal the outer block is too wide.

Catching KeyError on `d[key]` hides whether the key was missing and can mask typos or logic errors. Using `d.get(key)` or `if key in d:` makes the missing-key case explicit and keeps exception handling reserved for genuinely exceptional conditions.

Narrowing a broad except clause can surface errors that were previously silenced, changing control flow and potentially breaking callers that relied on swallowed failures. Always trace callees to know what they can raise and run the test suite after editing.