def test_iter():
one_to_ten = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
# f object support iteration on it's output
func = f(map) << 1 .__add__ << range(10) | list
assert func.value == one_to_ten
# func is an f object, not a list
assert type(func) == f != list
# f object is iterable
res = []
for n in func:
res.append(n)
assert res == one_to_ten
# works in list comprehension
res = [n for n in func]
assert res == one_to_ten
# and as arguments for functions that expect an iterator
res = list(func)
assert res == one_to_ten
# works with functions return a single non-iterable value, too
# f object will trun it into a 1-tuple
answer = f(42)
res = []
for n in answer:
res.append(n)
assert res == [42]
res = [n for n in answer]
assert res == [42]
res = list(answer)
assert res == [42]
def test_as_identity():
# f works as identity function on constant (non-callable)
assert f(1) == 1
assert f(2) != 1
# f wraps functions and works as original functions
assert f(sum(range(10))) == f(sum)(range(10)) == sum(f(range)(10)) == 45
def test_reverse_apply():
minus = f(lambda a, b: a - b)
assert minus(2, 1) == 1
assert minus.reverse_apply()(2, 1) == -1
# support more than 2 arguments
make_list = f(lambda *args: list(args))
make_reversed_list = make_list.reverse_apply()
assert make_list(1, 2, 3, 4, 5) == [1, 2, 3, 4, 5]
assert make_reversed_list(1, 2, 3, 4, 5) == [5, 4, 3, 2, 1]
def test_operator_pipe():
filter_ = f(filter)
map_ = f(map)
odd = lambda n: n % 2 == 1
snd = lambda e: e[1]
# as the name implies, it works like pipe in POSIX shell
odd_num = range(10) | filter_ << odd | enumerate | map_ << snd | list
assert odd_num == [1, 3, 5, 7, 9]
# pipe is equivalent to function composition in reversed order
fc = list ** (map_ << snd) ** enumerate ** (filter_ << odd) ** range(10)
assert fc == odd_num
def test_auto_eval():
# equality test on invokes function wrap by f object
# this is done by overloading operator == and != of f object
assert f(sum(range(1, 11))) == 55
# works in both directions
assert 55 == f(sum(range(1, 11)))
# even with f objects on both sides
assert f(sum(range(1, 11))) == f(sum(range(1, 11)))
g = f(sum(range(1, 101)))
assert g == g.value == g() == g.call() == g.invoke() == 5050
def test_operator_compose():
# make some curried functions
add_2 = add << 2
mul_5 = mul << 5
# and make sure they work
assert add_2(4) == 6
assert mul_5(3) == 15
# operator ** can be used for function composition
add_2_mul_5 = add_2 ** mul_5
assert add_2_mul_5(1) == 7
assert add_2_mul_5(2) == 12
assert add_2_mul_5(3) == 17
# pay attention to the order of evaluation, operator ** is
# right-associative, this is in accordance with the function composition
# operator in haskell, the dot operator (.)
mul_5_add_2 = mul_5 ** add_2
assert mul_5_add_2(1) == 15
assert mul_5_add_2(2) == 20
assert mul_5_add_2(3) == 25
# one more example of right-associative
assert neg ** abs << -1 == -1
# function composition in action
from itertools import count, takewhile as tw
takewhile = f(tw)
lt_20 = lambda n: n < 20
reverse = lambda s: s[::-1]
r = [19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0]
assert reverse ** list ** (takewhile << lt_20) ** count == r
# result of function composition will alway be f object, which means, in a
# chain of function composition, we can have as few as a well-placed
# single f object to make the expression valid. in order words, both
# __pow__ and __rpow__ are implemented.
assert f(list) ** f(map) << (add << 1) << range(10) == list(range(1, 11))
assert list ** f(map) << (add << 1) << range(10) == list(range(1, 11))
# to see it more clearly with identity function
identity = lambda x: x
# __pow__ in action
assert f(identity) ** 42 == 42
# __rpow__ in action
assert identity ** f(42) == 42
assert identity ** identity ** f(42) == 42
assert identity ** identity ** identity ** f(42) == 42
assert identity ** identity ** identity ** identity ** f(42) == 42
def test_contains():
# for function with single value
f_always_true = f(True)
# works like equality check
assert True in f_always_true
assert False not in f_always_true
# for function with sequence as value
one_to_ten = range(1, 11)
list_of_one_to_ten = f(list) << one_to_ten
# performs membership test on value
for i in one_to_ten:
assert i in list_of_one_to_ten
# make sure it did check membership instead of return True blindly
minus_one_to_ten = range(-1, -11, -1)
for i in minus_one_to_ten:
assert i not in list_of_one_to_ten
def test_clone():
# clone should always return a new instance.
the_answer = f(42)
assert the_answer is the_answer
assert the_answer is not the_answer.clone(the_answer)
assert the_answer is not the_answer.clone(42)
assert the_answer.clone(the_answer) is not the_answer.clone(42)
assert the_answer.clone(the_answer) is not the_answer.clone(the_answer)
assert the_answer.clone(42) is not the_answer.clone(42)
assert the_answer is the_answer
def test_flip():
# f.flip, as function flip in haskell
minus = f(lambda a, b: a - b)
assert minus << 3 << 2 == 1
two_minus = minus << 2
assert two_minus << 1 == 1
minus_two = minus.flip << 2
assert minus_two << 1 == -1
# flip twice yield the same function (sort of, there might be overhead
# added in reality)
assert minus(2, 1) == minus.flip.flip(2, 1)
def test_constructor():
# works with lambda expressions
assert f(lambda: 42) == 42
assert f(lambda a, b: a + b)(5, 6) == 11
# works with buildin functions
assert f(len)([1, 1, 2, 3, 5]) == 5
assert f(min)(3, 2, 1) == 1
# works with type constructors
assert f(set)() == set()
assert f(list)(range(3)) == [0, 1, 2]
def test_pipe():
one = f(1)
res = one.pipe(add << 2).pipe(mul << 5).pipe(sub.flip << 40).pipe(neg)
# -(40 - (1 + 2) * 5) == 25
assert res == 25
# equivalent operator pipe form
res = one | add << 2 | mul << 5 | sub.flip << 40 | neg
assert res == 25
g = (add << 2).pipe(mul << 5).pipe(sub.flip << 40).pipe(neg)
res = one.pipe(g)
assert res == 25
# equivalent operator pipe form
g = add << 2 | mul << 5 | sub.flip << 40 | neg
res = one | g
assert res == 25
def test_apply():
# applying arguments to f object returns partial applied functions.
# when apply one by one, it's called currying.
assert add.apply(1).apply(2)() == 3
assert sub.apply(1).apply(2)() == -1
# apply can accept more than one arguments at a time
add_1_4 = add.apply(1, 4)
assert add_1_4 == 5
# supports keyword arguments as well, notice named argument 'base' is
# supposed to be the second argument of int
from_hex = f(int).apply(base=16)
assert from_hex('0a') == 10
assert from_hex('0b') == 11
assert from_hex('0c') == 12
assert from_hex('0d') == 13
assert from_hex('0e') == 14
assert from_hex('0f') == 15
assert from_hex('10') == 16
def test_operator_high_cohesive_invoke():
# f objects can be called as functions
assert add(1, 2) == 3
assert sub(7, 3) == 4
assert mul(3, 7) == 21
assert truediv(32, 4) == 8
assert neg(1) == -1
# works with arbitrary numbers of arguments
assert f(bool)() is False
assert f(max)([1]) == 1
assert f(max)(6, 1) == 6
assert f(max)(6, 1, 9) == 9
assert f(max)(*range(10)) == 9
# works with keyword arguments
assert f(int)('10', base=16) == 16
def test_operator_high_cohesive_apply():
# operator << on f object is overloaded, works as f.apply
assert (add << 1 << 2) == 3
assert (sub << 1 << 2) == -1
# because of operator == has lower precedence than operator <<,
# parentheses around f object can be omitted.
assert add << 1 << 2 == 3
assert sub << 1 << 2 == -1
# add takes two arguments
assert add(1, 2) == 3
# apply one argument to it, make it a curried function that takes one
# argument
add_1 = add << 1
# invokes it with one argument yields result.
assert add_1(1) == 2
assert add_1(2) == 3
assert add_1(3) == 4
# curry again, becomes a function takes no argument
add_1_to_4 = add_1 << 4
assert add_1_to_4() == 5
# do it all in one expression
assert (add << 1 << 4)() == 5
# chaining apply
res = list(str(n) for n in range(10))
assert f(map) << str << range(10) | list == res
# augmented assignments supported, too
fm = f(map)
fm <<= str
fm <<= range(10)
assert list(fm.value) == res
# combine with other operators
list_of_numbers = ~f(map) << range(5) | list
add_1 = f(1 .__add__)
assert list_of_numbers << add_1 == [1, 2, 3, 4, 5]
times_2 = f(2 .__mul__)
assert list_of_numbers << times_2 == [0, 2, 4, 6, 8]
divided_by_2 = ~f(truediv) << 2
assert list_of_numbers << divided_by_2 == [0, 0.5, 1, 1.5, 2]
from fx.function import Function as f
from functools import partial
from operator import add, sub, mul, truediv, neg
# wrap them all with f objects
add, sub, mul, truediv, neg = f(add), f(sub), f(mul), f(truediv), f(neg)
def test_constructor():
# works with lambda expressions
assert f(lambda: 42) == 42
assert f(lambda a, b: a + b)(5, 6) == 11
# works with buildin functions
assert f(len)([1, 1, 2, 3, 5]) == 5
assert f(min)(3, 2, 1) == 1
# works with type constructors
assert f(set)() == set()
assert f(list)(range(3)) == [0, 1, 2]
def test_clone():
# clone should always return a new instance.
the_answer = f(42)
assert the_answer is the_answer
assert the_answer is not the_answer.clone(the_answer)
assert the_answer is not the_answer.clone(42)
assert the_answer.clone(the_answer) is not the_answer.clone(42)
assert the_answer.clone(the_answer) is not the_answer.clone(the_answer)
assert the_answer.clone(42) is not the_answer.clone(42)
def test_operator_flip():
minus = f(lambda a, b: a - b)
# operator ~ works as f.flip
minus_two = ~minus << 2
assert minus_two << 1 == -1
def test_operator_low_cohesive_apply():
# when using higher-order functions, parentheses might be needed
assert f(map) << (add << 1) << range(5) | list == [1, 2, 3, 4, 5]
# with low cohesive apply operator, parentheses can be omitted
assert f(map) & add << 1 & range(5) | list == [1, 2, 3, 4, 5]
def test_partial_compatible():
# f object works with partial functions (created by functools.partial)
add = lambda a, b: a + b
add_1 = partial(add, 1)
assert add_1(2) == 3
assert f(add_1) << 2 == 3