What is the explanation for length of a block? - ruby

I am wondering about this line:
{ |i| (num % i) == 0}.length == 0
Can someone explain what happens using length on a block? I am trying to answer a rubeque problem about prime factors:
divisors = Array.new
for d in 2..(num)
divisors << d if (num % d) == 0
end
primes = divisors.select do |num|
(2..(num-1)).select { |i| (num % i) == 0}.length == 0
end

It's calling length on the result of (2..(num-1)).select { |i| (num % i) == 0} — which is an array — not on the block.

Related

Is there a known algorithm for simplifying a boolean expression with number comparisons?

For example, if I have the expression (A > 5) && (A == 6),
that expression can be simplified to just (A == 6), and still have the same behavior for A ∈ ℤ.
I also need it to work with multiple variables, so for instance ((B > 2) && (C == 2)) || ((B > 2) && (C < 2)) should simplify to (B > 2) && (C < 3).
I won't need to compare two unknowns, only unknowns and numbers, and I only need it to work with the operators <, >, and == for numbers, and && and || for expressions (&& being AND and || being OR, of course). All unknowns are integers.
Is there any algorithm that takes such an expression and returns an expression with equal behavior and a minimal amount of operators?
(in my specific case, || operators are preferred over &&)
Here's a slow dynamic programming algorithm along the lines that you were thinking of.
from collections import defaultdict, namedtuple
from heapq import heappop, heappush
from itertools import product
from math import inf
# Constructors for Boolean expressions. False and True are also accepted.
Lt = namedtuple("Lt", ["lhs", "rhs"])
Eq = namedtuple("Eq", ["lhs", "rhs"])
Gt = namedtuple("Gt", ["lhs", "rhs"])
And = namedtuple("And", ["lhs", "rhs"])
Or = namedtuple("Or", ["lhs", "rhs"])
# Variable names. Arbitrary strings are accepted.
A = "A"
B = "B"
C = "C"
# Example formulas.
first_example = And(Gt(A, 5), Eq(A, 6))
second_example = Or(And(Gt(B, 2), Eq(C, 2)), And(Gt(B, 2), Lt(C, 2)))
third_example = Or(And(Gt(A, 1), Gt(B, 1)), And(Gt(A, 0), Gt(B, 2)))
fourth_example = Or(Lt(A, 6), Gt(A, 5))
fifth_example = Or(And(Eq(A, 2), Gt(C, 2)), And(Eq(B, 2), Lt(C, 2)))
# Returns a map from each variable to the set of values such that the formula
# might evaluate differently for variable = value-1 versus variable = value.
def get_critical_value_sets(formula, result=None):
if result is None:
result = defaultdict(set)
if isinstance(formula, bool):
pass
elif isinstance(formula, Lt):
result[formula.lhs].add(formula.rhs)
elif isinstance(formula, Eq):
result[formula.lhs].add(formula.rhs)
result[formula.lhs].add(formula.rhs + 1)
elif isinstance(formula, Gt):
result[formula.lhs].add(formula.rhs + 1)
elif isinstance(formula, (And, Or)):
get_critical_value_sets(formula.lhs, result)
get_critical_value_sets(formula.rhs, result)
else:
assert False, str(formula)
return result
# Returns a list of inputs sufficient to compare Boolean combinations of the
# primitives returned by enumerate_useful_primitives.
def enumerate_truth_table_inputs(critical_value_sets):
variables, value_sets = zip(*critical_value_sets.items())
return [
dict(zip(variables, values))
for values in product(*({-inf} | value_set for value_set in value_sets))
]
# Returns both constants and all single comparisons whose critical value set is
# a subset of the given ones.
def enumerate_useful_primitives(critical_value_sets):
yield False
yield True
for variable, value_set in critical_value_sets.items():
for value in value_set:
yield Lt(variable, value)
if value + 1 in value_set:
yield Eq(variable, value)
yield Gt(variable, value - 1)
# Evaluates the formula recursively on the given input.
def evaluate(formula, input):
if isinstance(formula, bool):
return formula
elif isinstance(formula, Lt):
return input[formula.lhs] < formula.rhs
elif isinstance(formula, Eq):
return input[formula.lhs] == formula.rhs
elif isinstance(formula, Gt):
return input[formula.lhs] > formula.rhs
elif isinstance(formula, And):
return evaluate(formula.lhs, input) and evaluate(formula.rhs, input)
elif isinstance(formula, Or):
return evaluate(formula.lhs, input) or evaluate(formula.rhs, input)
else:
assert False, str(formula)
# Evaluates the formula on the many inputs, packing the values into an integer.
def get_truth_table(formula, inputs):
truth_table = 0
for input in inputs:
truth_table = (truth_table << 1) + evaluate(formula, input)
return truth_table
# Returns (the number of operations in the formula, the number of Ands).
def get_complexity(formula):
if isinstance(formula, bool):
return (0, 0)
elif isinstance(formula, (Lt, Eq, Gt)):
return (1, 0)
elif isinstance(formula, And):
ops_lhs, ands_lhs = get_complexity(formula.lhs)
ops_rhs, ands_rhs = get_complexity(formula.rhs)
return (ops_lhs + 1 + ops_rhs, ands_lhs + 1 + ands_rhs)
elif isinstance(formula, Or):
ops_lhs, ands_lhs = get_complexity(formula.lhs)
ops_rhs, ands_rhs = get_complexity(formula.rhs)
return (ops_lhs + 1 + ops_rhs, ands_lhs + ands_rhs)
else:
assert False, str(formula)
# Formula compared by complexity.
class HeapItem:
__slots__ = ["_complexity", "formula"]
def __init__(self, formula):
self._complexity = get_complexity(formula)
self.formula = formula
def __lt__(self, other):
return self._complexity < other._complexity
def __le__(self, other):
return self._complexity <= other._complexity
def __eq__(self, other):
return self._complexity == other._complexity
def __ne__(self, other):
return self._complexity != other._complexity
def __ge__(self, other):
return self._complexity >= other._complexity
def __gt__(self, other):
return self._complexity > other._complexity
# Like heapq.merge except we can add iterables dynamically.
class Merge:
__slots__ = ["_heap", "_iterable_count"]
def __init__(self):
self._heap = []
self._iterable_count = 0
def update(self, iterable):
iterable = iter(iterable)
try:
value = next(iterable)
except StopIteration:
return
heappush(self._heap, (value, self._iterable_count, iterable))
self._iterable_count += 1
def __iter__(self):
return self
def __next__(self):
if not self._heap:
raise StopIteration
value, index, iterable = heappop(self._heap)
try:
next_value = next(iterable)
except StopIteration:
return value
heappush(self._heap, (next_value, index, iterable))
return value
class Combinations:
__slots__ = ["_op", "_formula", "_best_formulas", "_i", "_n"]
def __init__(self, op, formula, best_formulas):
self._op = op
self._formula = formula
self._best_formulas = best_formulas
self._i = 0
self._n = len(best_formulas)
def __iter__(self):
return self
def __next__(self):
if self._i >= self._n:
raise StopIteration
formula = self._op(self._formula, self._best_formulas[self._i])
self._i += 1
return HeapItem(formula)
# Returns the simplest equivalent formula, breaking ties in favor of fewer Ands.
def simplify(target_formula):
critical_value_sets = get_critical_value_sets(target_formula)
inputs = enumerate_truth_table_inputs(critical_value_sets)
target_truth_table = get_truth_table(target_formula, inputs)
best = {}
merge = Merge()
for formula in enumerate_useful_primitives(critical_value_sets):
merge.update([HeapItem(formula)])
best_formulas = []
for item in merge:
if target_truth_table in best:
return best[target_truth_table]
formula = item.formula
truth_table = get_truth_table(formula, inputs)
if truth_table in best:
continue
n = len(best_formulas)
for op in [And, Or]:
merge.update(Combinations(op, formula, best_formulas))
best[truth_table] = formula
best_formulas.append(formula)
print(simplify(first_example))
print(simplify(second_example))
print(simplify(third_example))
print(simplify(fourth_example))
print(simplify(fifth_example))
Output:
Eq(lhs='A', rhs=6)
And(lhs=Lt(lhs='C', rhs=3), rhs=Gt(lhs='B', rhs=2))
And(lhs=And(lhs=Gt(lhs='B', rhs=1), rhs=Gt(lhs='A', rhs=0)), rhs=Or(lhs=Gt(lhs='B', rhs=2), rhs=Gt(lhs='A', rhs=1)))
True
Or(lhs=And(lhs=Eq(lhs='B', rhs=2), rhs=Lt(lhs='C', rhs=2)), rhs=And(lhs=Gt(lhs='C', rhs=2), rhs=Eq(lhs='A', rhs=2)))
Maybe you can consider intervals for your variables, for example:
(A > 5) && (A == 6)
Given you have a variable A, set an initial interval for it: A: [-∞, ∞].
Each condition that you read, you can reduce your interval:
(A > 5) sets the interval for A: [6, ∞]
(A == 6) sets the interval for A: [6, 6]
For each update on the interval, check if the new condition is possible, for example:
(A > 5) sets the interval for A: [6, ∞]
(A == 5) out of the interval, impossible condition.
Just another example:
((B > 2) && (C == 2)) || ((B > 2) && (C < 2))
Initially: B: [-∞, ∞] and C: [-∞, ∞].
((B > 2) && (C == 2))
(B > 2) sets the interval for B: [3, ∞]
(C == 2) sets the interval for C: [2, 2]
The next condition is attached with ||, so you add intervals:
((B > 2) && (C < 2))
(B > 2) sets the interval for B: [3, ∞]
(C < 2) sets the interval for C: [2, 2] U [-∞, 1] = [-∞, 2]

How to optimize code - it works, but I know I'm missing much learning

The exercise I'm working on asks "Write a method, coprime?(num_1, num_2), that accepts two numbers as args. The method should return true if the only common divisor between the two numbers is 1."
I've written a method to complete the task, first by finding all the factors then sorting them and looking for duplicates. But I'm looking for suggestions on areas I should consider to optimize it.
The code works, but it is just not clean.
def factors(num)
return (1..num).select { |n| num % n == 0}
end
def coprime?(num_1, num_2)
num_1_factors = factors(num_1)
num_2_factors = factors(num_2)
all_factors = num_1_factors + num_2_factors
new = all_factors.sort
dups = 0
new.each_index do |i|
dups += 1 if new[i] == new[i+1]
end
if dups > 1
false
else
true
end
end
p coprime?(25, 12) # => true
p coprime?(7, 11) # => true
p coprime?(30, 9) # => false
p coprime?(6, 24) # => false
You could use Euclid's algorithm to find the GCD, then check whether it's 1.
def gcd a, b
while a % b != 0
a, b = b, a % b
end
return b
end
def coprime? a, b
gcd(a, b) == 1
end
p coprime?(25, 12) # => true
p coprime?(7, 11) # => true
p coprime?(30, 9) # => false
p coprime?(6, 24) # => false```
You can just use Integer#gcd:
def coprime?(num_1, num_2)
num_1.gcd(num_2) == 1
end
You don't need to compare all the factors, just the prime ones. Ruby does come with a Prime class
require 'prime'
def prime_numbers(num_1, num_2)
Prime.each([num_1, num_2].max / 2).map(&:itself)
end
def factors(num, prime_numbers)
prime_numbers.select {|n| num % n == 0}
end
def coprime?(num_1, num_2)
prime_numbers = prime_numbers(num_1, num_2)
# & returns the intersection of 2 arrays (https://stackoverflow.com/a/5678143)
(factors(num_1, prime_numbers) & factors(num_2, prime_numbers)).length == 0
end

Ruby FizzBuzz using arrays, my logic seems right but it is getting an error

FizzBuzz, a classic problem, returns all the numbers up to N with a slight twist. If a number is divisible by 3, it is replaced with "fizz". If it's divisible by 5, it's replaced with "buzz". If it's divisible by both, it's replaced with "fizzbuzz"
I keep getting this Error message:
comparison of Fixnum with nil failed
Can someone explain this error message to me please? Also why is the code not working?
def fizz_buzz(n)
arr = (1..n).to_a
i = 0
while arr[i] < arr[n]
if i % 3 == 0 && i % 5 == 0
arr[i] = 'fizzbuzz'
elsif i % 3 == 0
arr[i] = 'fizz'
elsif i % 5 == 0
arr[i] = 'buzz'
else
arr[i] = i
i += 1
end
end
return arr
end
fizz_buzz(12)
Your conditions are just a bit off, give this a try:
def fizz_buzz(n)
arr = (1..n).to_a
i = 0
while i < n
if arr[i] % 3 == 0 && arr[i] % 5 == 0
arr[i] = 'fizzbuzz'
elsif arr[i] % 3 == 0
arr[i] = 'fizz'
elsif arr[i] % 5 == 0
arr[i] = 'buzz'
end
i+=1
end
return arr
end
Trying to access arr[n] puts you outside the bounds of the array which returns nil in Ruby.
You can update the code the ruby way, by using blocks and guards, I don't remember last time I used a while loop in ruby :)
Also, Array.new accepts a block as an argument which you can exploit to build your Array in a single step:
def fizz_buzz(n)
Array.new(n) do |index|
x = index + 1
case
when x % 3 == 0 && x % 5 == 0 then "fizzbuzz"
when x % 3 == 0 then "fizz"
when x % 5 == 0 then "buzz"
else x
end
end
end
Notice I used 1 as a base index and not 0, you can just remove x = index + 1 and replace x with index to have it working in a zero index base
A solution with a block instead of the while loop, and guards
def fizz_buzz(n)
arr = (1..n).to_a
0.upto(n - 1) do |i|
arr[i] = "fizzbuzz" and next if i % 3 == 0 && i % 5 == 0
arr[i] = "fizz" and next if i % 3 == 0
arr[i] = "buzz" if i % 5 == 0
end
arr
end
#brad-melanson beat me to the straight-forward answer to your question, so I'll share an answer which uses some common Ruby idioms (passing a range to the Array constructor and map), which simplify things, prevent you from having to do any iteration bookkeeping and prevent the possibility of off-by-one errors, out-of-bounds errors, etc.
def fizz_buzz(n)
Array(1..12).map do |n|
if n % 3 == 0 && n % 5 == 0
'fizzbuzz'
elsif n % 3 == 0
'fizz'
elsif n % 5 == 0
'buzz'
else
n
end
end
end
result = fizz_buzz 12
# result => [1, 2, "fizz", 4, "buzz", "fizz", 7, 8, "fizz", "buzz", 11, "fizz"]

Project Euler #3 Ruby

The task:
The prime factors of 13195 are 5, 7, 13 and 29.
What is the largest prime factor of the number 600851475143?
The correct answer is 6857.
My code:
def prime?(n)
(2..(n-1)).each { |x| false if n % x == 0 }
true
end
x = 2
prime_factor_arr = []
number = 600_851_475_143
while x < number
if number % x == 0 && prime?(x)
prime_factor_arr << x
number = number / x
end
x += 1
end
puts prime_factor_arr.last
puts prime?(prime_factor_arr.last)
puts prime_factor_arr
In the above case, I get 1471 as the largest prime. If I change the code to:
while x < (number / x)
if number % x == 0 && prime?(x)
prime_factor_arr << x
end
x += 1
end
I get 486847. The array printed in the end is:
[71, 839, 1471, 6857, 59569, 104441, 486847]
It is not clear to me why my code does not work. Could anybody help?
For the answer to the question, Sergio is right. But the code Sergio suggests (as well as yours) will not work correctly when number is a prime itself.
A better way to write is:
def prime?(n); (2...n).none?{|x| n.%(x).zero?} end
number = 600_851_475_143
number.downto(1).find{|x| number.%(x).zero? and prime?(x)}
Try this.
def prime? n
(2..(n-1)).each { |x| return false if n % x == 0 }
true
end
n = 600_851_475_143
a = []
product_sum = 1
x = 2 # 2 is the first prime number
while product_sum < n
if n % x == 0 && prime?(x)
a << x
product_sum *= x
end
x += 1
end
puts "The answer is #{a.last}"
The prime lib (from standard lib) is very nice for Project Euler. But it takes the fun out of this one:
require "prime"
600851475143.prime_division.last.first # => 6857

Can I have a block in begin/while statement in Ruby?

I'm trying out to have a block in a while and begin statements in Ruby, but I get a syntax error. Any other way to implement it?
Here's what I want to accomplish
(1..limit).each { |i|
while (true) do |n|
x = n * (i%n)
puts n if n%i != 0
break if x.even? && !x.zero?
n += 1
end
}
while is a keyword, so you do not need the block. Your code should be:
(1..limit).each { |i|
while (true)
x = n * (i%n)
puts n if n%i != 0
break if x.even? && !x.zero?
n += 1
end
}
But you are requesting a block variable form the while statement. Variable names inside the pipes are for variables passed to your block containing information from whatever calls your block. I will assume that n is supposed to increment. Here is a working version:
(1..limit).each { |i|
n = 0
while (true)
x = n * (i%n)
puts n if n%i != 0
break if x.even? && !x.zero?
n += 1
end
}
If you really need the code in a block, you could create one and then call it, like this (ruby 1.9 only):
(1..limit).each { |i|
n = 0
while (true)
-> do
x = n * (i%n)
puts n if n%i != 0
break if x.even? && !x.zero?
n += 1
end.()
end
}
By the way, here is a cleaner version:
(1..limit).each do |i|
n = 0
loop do
x = n * (i % n)
puts n if n % i != 0
break if x.even? and !x.zero?
n += 1
end
end

Resources