DenkzeitWiki

Python Function Annotations

• PEP 3107: Function Annotations
  • Implementation
  • annotation does not have a standardized semantic - semantic must\can be added by 3rd party libraries
    
  
  
• also of interest: JavaAnnotations?

def haul(item: Haulable, *vargs: PackAnimal) -> Distance:
    ...

Use Cases

• OptionalStaticTyping (not really since no compile-time checks per se)
  • supporting IDEs (e.g. auto-completion)
  • type-annotation can be quite narrow (depending on 3rd-party-implementation), e.g. "non-negative integers" or "strings of max-length 13".
    • conventional types therefore sometimes even lie (or deceive the actual type)
• Adaptation (compare GenericFunctions)

for simple type-based dispatch, an inline way of specifying type or interface would be quite nice for generic functions.[1]

• Overloading

 
 @overload(int, str)
 def foo(bar, baz):
     ...

 @overload
 def foo(bar: int, baz: str)
     ...
 

• interfacing with statically typed languages

In other words, what I'm saying is that declarations could be a purely syntactic/metadata feature, and they'd still be quite useful for many things.

• the annotations must be introspectable (i.e. accessible in code, from outside the function)

Type Declaration

• http://osteele.com/archives/2004/12/type-declaration-compromise
  • http://osteele.com/archives/2005/01/three-lefts
  • OptionalStaticTyping (don't force type-declarations)

Discussion

• [Python-3000] What's the point of annotations?
• PEP 3107 and stronger typing (note: probably a newbie question)
• http://oakwinter.com/code/category/python/function-annotations-pep-3107/