Sort of like pipes, except that we declare abstract relations between functions.
Thinking of this as a combination extract function and extract variable refactor forms.
A few different reasons.
Python is dynamically typed. That's great.
Some data doesn't fit well within a typing system anyways.
For instance, a US based telephone number is a value based paradigm for a str
type. You could wrap it in a type or a class, but that might be a bit heavy-handed.
And maybe, you want to accept any kind of phone number, but then be able to query
which kind of phone number it could be, at run time. This is a style of typing
commonly seen in Lisp-derived languages where every type comes with an instantiator,
e.g. pair and a type tester, e.g. pair?.
And also, when it comes to regexes, it's as though I'm doing it for the very first time every time. It's one of those things that never ends up being DRY. Instead of creating another new regex, wouldn't it make more sense to create a value-testing function that could be reused in multiple code bases?
This library is meant to strongly emphasize "extract" type refactors and functional composition.
There are a few different ways to "misuse" this paradigm (IMO):
- Collapse tube types into overly dense, unreadable functional code
- TODO: link to
onelinerizer
- Port in BASH shell like paradigms to Python
See examples.py (for now).