A cool tool for functional programming. Operating Functions as in math. Type testing is implemented in an awsome way as well as memoizing.
Functional Programming, Type Testing
Feature:
1. function with domain, the domain can be used to type testing
2. operator on functions, for example, f + g: x -> f(x) + g(x) where f, g, f+g are functions
3. to glue functions with glue function or method that can be used to define piecewise functions
4. no 3rd part requirement
5. implement memoization in oo way.
6. use MathFunction for math functions where keyword arguments are deprecated. (new in this version)
Classes
BaseFunction -> Type (or Domain), Function
BaseFunction: func: function (or number)
Function: func,
domain: Type, the definition domain of func
Functions
Interval(a:num, b:num) -> Type
restrict(t:Type) -> decorator(f:function -> Function(f, t))
Constants
TURE, FALSE(Type) represent universal set, empty set
import:
import fcool (or from fcool import *)
Define Function with domain:
F = Function(lambda x:x, Type(lambda x:x>2))
F(3)
Operators on Functions:
(F + F)(3)
(F * F)(4)
(2 * F)(3)
2D Functions and Types:
f = 3 # or lambda x,y: 3
g = lambda x,y: 2/x
t = Type(lambda x:x<5 and isinstance(x, int)) * TRUE # define type(domain) and functions on it
<=> Type(lambda x, y:x<5) & Type(lambda x, y:isinstance(x, int))
G = Function(g, t)
F = Function(f, t)
Memoize (the coolest feature of the new version):
f.memoize() # f is the object of BaseFunction, similar to toolz.memoize(f)
f.unmemoize() # prohibit to use memo (memo is not deleted)
f.del_memo() # just clear the memo, will update the memo in next time
f.forget() # f.del_memo() and f.unmemoize()
Glue Functions:
ID = Function(lambda x:x)
print(ID.compose(F)(3,4)) # composition
Type testing with restrict decorator:
@restrict(Interval(1,2)) # restriction decorator
def f(x):
return x
print(f(1))
try:
print(f(3))
except Exception as ex:
print(ex)
G=Function(lambda x:x)
print(G(3))
G = G | Interval(1,2) # restriction method
print(G(3))