I know there is nothing wrong with writing with proper function structure, but I would like to know how can I find nth fibonacci number with most Pythonic way with a one-line.
I wrote that code, but It didn't seem to me best way:
>>> fib = lambda n:reduce(lambda x, y: (x[0]+x[1], x[0]), [(1,1)]*(n-2))[0]
>>> fib(8)
13
How could it be better and simplier?
fib = lambda n:reduce(lambda x,n:[x[1],x[0]+x[1]], range(n),[0,1])[0]
(this maintains a tuple mapped from [a,b] to [b,a+b], initialized to [0,1], iterated N times, then takes the first tuple element)
>>> fib(1000)
43466557686937456435688527675040625802564660517371780402481729089536555417949051
89040387984007925516929592259308032263477520968962323987332247116164299644090653
3187938298969649928516003704476137795166849228875L
(note that in this numbering, fib(0) = 0, fib(1) = 1, fib(2) = 1, fib(3) = 2, etc.)
(also note: reduce is a builtin in Python 2.7 but not in Python 3; you'd need to execute from functools import reduce in Python 3.)
A rarely seen trick is that a lambda function can refer to itself recursively:
fib = lambda n: n if n < 2 else fib(n-1) + fib(n-2)
By the way, it's rarely seen because it's confusing, and in this case it is also inefficient. It's much better to write it on multiple lines:
def fibs():
a = 0
b = 1
while True:
yield a
a, b = b, a + b
I recently learned about using matrix multiplication to generate Fibonacci numbers, which was pretty cool. You take a base matrix:
[1, 1]
[1, 0]
and multiply it by itself N times to get:
[F(N+1), F(N)]
[F(N), F(N-1)]
This morning, doodling in the steam on the shower wall, I realized that you could cut the running time in half by starting with the second matrix, and multiplying it by itself N/2 times, then using N to pick an index from the first row/column.
With a little squeezing, I got it down to one line:
import numpy
def mm_fib(n):
return (numpy.matrix([[2,1],[1,1]])**(n//2))[0,(n+1)%2]
>>> [mm_fib(i) for i in range(20)]
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, 4181]
This is a closed expression for the Fibonacci series that uses integer arithmetic, and is quite efficient.
fib = lambda n:pow(2<<n,n+1,(4<<2*n)-(2<<n)-1)%(2<<n)
>> fib(1000)
4346655768693745643568852767504062580256466051737178
0402481729089536555417949051890403879840079255169295
9225930803226347752096896232398733224711616429964409
06533187938298969649928516003704476137795166849228875L
It computes the result in O(log n) arithmetic operations, each acting on integers with O(n) bits. Given that the result (the nth Fibonacci number) is O(n) bits, the method is quite reasonable.
It's based on genefib4 from http://fare.tunes.org/files/fun/fibonacci.lisp , which in turn was based on an a less efficient closed-form integer expression of mine (see: http://paulhankin.github.io/Fibonacci/)
If we consider the "most Pythonic way" to be elegant and effective then:
def fib(nr):
return int(((1 + math.sqrt(5)) / 2) ** nr / math.sqrt(5) + 0.5)
wins hands down. Why use a inefficient algorithm (and if you start using memoization we can forget about the oneliner) when you can solve the problem just fine in O(1) by approximation the result with the golden ratio? Though in reality I'd obviously write it in this form:
def fib(nr):
ratio = (1 + math.sqrt(5)) / 2
return int(ratio ** nr / math.sqrt(5) + 0.5)
More efficient and much easier to understand.
This is a non-recursive (anonymous) memoizing one liner
fib = lambda x,y=[1,1]:([(y.append(y[-1]+y[-2]),y[-1])[1] for i in range(1+x-len(y))],y[x])[1]
fib = lambda n, x=0, y=1 : x if not n else fib(n-1, y, x+y)
run time O(n), fib(0) = 0, fib(1) = 1, fib(2) = 1 ...
I'm Python newcomer, but did some measure for learning purposes. I've collected some fibo algorithm and took some measure.
from datetime import datetime
import matplotlib.pyplot as plt
from functools import wraps
from functools import reduce
from functools import lru_cache
import numpy
def time_it(f):
#wraps(f)
def wrapper(*args, **kwargs):
start_time = datetime.now()
f(*args, **kwargs)
end_time = datetime.now()
elapsed = end_time - start_time
elapsed = elapsed.microseconds
return elapsed
return wrapper
#time_it
def fibslow(n):
if n <= 1:
return n
else:
return fibslow(n-1) + fibslow(n-2)
#time_it
#lru_cache(maxsize=10)
def fibslow_2(n):
if n <= 1:
return n
else:
return fibslow_2(n-1) + fibslow_2(n-2)
#time_it
def fibfast(n):
if n <= 1:
return n
a, b = 0, 1
for i in range(1, n+1):
a, b = b, a + b
return a
#time_it
def fib_reduce(n):
return reduce(lambda x, n: [x[1], x[0]+x[1]], range(n), [0, 1])[0]
#time_it
def mm_fib(n):
return (numpy.matrix([[2, 1], [1, 1]])**(n//2))[0, (n+1) % 2]
#time_it
def fib_ia(n):
return pow(2 << n, n+1, (4 << 2 * n) - (2 << n)-1) % (2 << n)
if __name__ == '__main__':
X = range(1, 200)
# fibslow_times = [fibslow(i) for i in X]
fibslow_2_times = [fibslow_2(i) for i in X]
fibfast_times = [fibfast(i) for i in X]
fib_reduce_times = [fib_reduce(i) for i in X]
fib_mm_times = [mm_fib(i) for i in X]
fib_ia_times = [fib_ia(i) for i in X]
# print(fibslow_times)
# print(fibfast_times)
# print(fib_reduce_times)
plt.figure()
# plt.plot(X, fibslow_times, label='Slow Fib')
plt.plot(X, fibslow_2_times, label='Slow Fib w cache')
plt.plot(X, fibfast_times, label='Fast Fib')
plt.plot(X, fib_reduce_times, label='Reduce Fib')
plt.plot(X, fib_mm_times, label='Numpy Fib')
plt.plot(X, fib_ia_times, label='Fib ia')
plt.xlabel('n')
plt.ylabel('time (microseconds)')
plt.legend()
plt.show()
The result is usually the same.
Fiboslow_2 with recursion and cache, Fib integer arithmetic and Fibfast algorithms seems the best ones. Maybe my decorator not the best thing to measure performance, but for an overview it seemed good.
Another example, taking the cue from Mark Byers's answer:
fib = lambda n,a=0,b=1: a if n<=0 else fib(n-1,b,a+b)
I wanted to see if I could create an entire sequence, not just the final value.
The following will generate a list of length 100. It excludes the leading [0, 1] and works for both Python2 and Python3. No other lines besides the one!
(lambda i, x=[0,1]: [(x.append(x[y+1]+x[y]), x[y+1]+x[y])[1] for y in range(i)])(100)
Output
[1,
2,
3,
...
218922995834555169026,
354224848179261915075,
573147844013817084101]
Here's an implementation that doesn't use recursion, and only memoizes the last two values instead of the whole sequence history.
nthfib() below is the direct solution to the original problem (as long as imports are allowed)
It's less elegant than using the Reduce methods above, but, although slightly different that what was asked for, it gains the ability to to be used more efficiently as an infinite generator if one needs to output the sequence up to the nth number as well (re-writing slightly as fibgen() below).
from itertools import imap, islice, repeat
nthfib = lambda n: next(islice((lambda x=[0, 1]: imap((lambda x: (lambda setx=x.__setitem__, x0_temp=x[0]: (x[1], setx(0, x[1]), setx(1, x0_temp+x[1]))[0])()), repeat(x)))(), n-1, None))
>>> nthfib(1000)
43466557686937456435688527675040625802564660517371780402481729089536555417949051
89040387984007925516929592259308032263477520968962323987332247116164299644090653
3187938298969649928516003704476137795166849228875L
from itertools import imap, islice, repeat
fibgen = lambda:(lambda x=[0,1]: imap((lambda x: (lambda setx=x.__setitem__, x0_temp=x[0]: (x[1], setx(0, x[1]), setx(1, x0_temp+x[1]))[0])()), repeat(x)))()
>>> list(islice(fibgen(),12))
[1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144]
def fib(n):
x =[0,1]
for i in range(n):
x=[x[1],x[0]+x[1]]
return x[0]
take the cue from Jason S, i think my version have a better understanding.
Starting Python 3.8, and the introduction of assignment expressions (PEP 572) (:= operator), we can use and update a variable within a list comprehension:
fib = lambda n,x=(0,1):[x := (x[1], sum(x)) for i in range(n+1)][-1][0]
This:
Initiates the duo n-1 and n-2 as a tuple x=(0, 1)
As part of a list comprehension looping n times, x is updated via an assignment expression (x := (x[1], sum(x))) to the new n-1 and n-2 values
Finally, we return from the last iteration, the first part of the x
To solve this problem I got inspired by a similar question here in Stackoverflow Single Statement Fibonacci, and I got this single line function that can output a list of fibonacci sequence. Though, this is a Python 2 script, not tested on Python 3:
(lambda n, fib=[0,1]: fib[:n]+[fib.append(fib[-1] + fib[-2]) or fib[-1] for i in range(n-len(fib))])(10)
assign this lambda function to a variable to reuse it:
fib = (lambda n, fib=[0,1]: fib[:n]+[fib.append(fib[-1] + fib[-2]) or fib[-1] for i in range(n-len(fib))])
fib(10)
output is a list of fibonacci sequence:
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]
I don't know if this is the most pythonic method but this is the best i could come up with:->
Fibonacci = lambda x,y=[1,1]:[1]*x if (x<2) else ([y.append(y[q-1] + y[q-2]) for q in range(2,x)],y)[1]
The above code doesn't use recursion, just a list to store the values.
My 2 cents
# One Liner
def nthfibonacci(n):
return long(((((1+5**.5)/2)**n)-(((1-5**.5)/2)**n))/5**.5)
OR
# Steps
def nthfibonacci(nth):
sq5 = 5**.5
phi1 = (1+sq5)/2
phi2 = -1 * (phi1 -1)
n1 = phi1**(nth+1)
n2 = phi2**(nth+1)
return long((n1 - n2)/sq5)
Why not use a list comprehension?
from math import sqrt, floor
[floor(((1+sqrt(5))**n-(1-sqrt(5))**n)/(2**n*sqrt(5))) for n in range(100)]
Without math imports, but less pretty:
[int(((1+(5**0.5))**n-(1-(5**0.5))**n)/(2**n*(5**0.5))) for n in range(100)]
import math
sqrt_five = math.sqrt(5)
phi = (1 + sqrt_five) / 2
fib = lambda n : int(round(pow(phi, n) / sqrt_five))
print([fib(i) for i in range(1, 26)])
single line lambda fibonacci but with some extra variables
Similar:
def fibonacci(n):
f=[1]+[0]
for i in range(n):
f=[sum(f)] + f[:-1]
print f[1]
A simple Fibonacci number generator using recursion
fib = lambda x: 1-x if x < 2 else fib(x-1)+fib(x-2)
print fib(100)
This takes forever to calculate fib(100) in my computer.
There is also closed form of Fibonacci numbers.
fib = lambda n: int(1/sqrt(5)*((1+sqrt(5))**n-(1-sqrt(5))**n)/2**n)
print fib(50)
This works nearly up to 72 numbers due to precision problem.
Lambda with logical operators
fibonacci_oneline = lambda n = 10, out = []: [ out.append(i) or i if i <= 1 else out.append(out[-1] + out[-2]) or out[-1] for i in range(n)]
here is how i do it ,however the function returns None for the list comprehension line part to allow me to insert a loop inside ..
so basically what it does is appending new elements of the fib seq inside of a list which is over two elements
>>f=lambda list,x :print('The list must be of 2 or more') if len(list)<2 else [list.append(list[-1]+list[-2]) for i in range(x)]
>>a=[1,2]
>>f(a,7)
You can generate once a list with some values and use as needed:
fib_fix = []
fib = lambda x: 1 if x <=2 else fib_fix[x-3] if x-2 <= len(fib_fix) else (fib_fix.append(fib(x-2) + fib(x-1)) or fib_fix[-1])
fib_x = lambda x: [fib(n) for n in range(1,x+1)]
fib_100 = fib_x(100)
than for example:
a = fib_fix[76]
Related
I'm writing a program that evaluates the power series sum_{m=0}{oo} a[m]x^m, where a[m] is recursively defined: a[m]=f(a[m-1]). I am generating symbols as follows:
a = list(sympy.symbols(' '.join([('a%d' % i) for i in range(10)])))
for i in range(1, LIMIT):
a[i] = f_recur(a[i-1], i-1)
This lets me refer to the symbols a0,a1,...,a9 using a[0],a[1],...,a[9], and a[m] is a function of a[m-1] given by f_recur.
Now, I hope code up the summation as follows:
m, x, y = sympy.symbols('m x y')
y = sympy.Sum(a[m]*x**m, (m, 0, 10))
But, m is not an integer so a[m] throws an Exception.
In this situation, where symbols are stored in a list, how would you code the summation? Thanks for any help!
SymPy's Sum is designed as a sum with a symbolic index. You want a sum with a concrete index running through 0, ... 9. This could be Python's sum
y = sum([a[m]*x**m for m in range(10)])
or, which is preferable from the performance point of view (relevant issue)
y = sympy.Add(*[a[m]*x**m for m in range(10)])
In either case, m is not a symbol but an integer.
I have a work-around that does not use sympy.Sum:
x = sympy.symbols('x')
y = a[0]*x**0
for i in range(1, LIMIT):
y += a[i]*x**i
This does the job, but sympy.Sum is not used.
Use IndexedBase instead of Symbol:
>>> a = IndexedBase('a')
>>> Sum(x**m*a[m],(m,1,3))
Sum(a[m]*x**m, (m, 1, 3))
>>> _.doit()
a[1]*x + a[2]*x**2 + a[3]*x**3
Suppose we have a defined function as following, and we would like to iterate over n from 1 to L, I've suffered a lot for a vectorization code, since this code is rather slow due to for loop needed outside to call this function.
Details: L, K are large integers e.g. 1000 and H_n is float value.
def multifrac_Brownian_motion(n, L, K, list_hurst, ind_hurst):
t_ks = np.asarray(sorted(-np.array(range(1, K + 1))*(1./L)))
t_ns = np.linspace(0, 1, num=L+1)
t_n = t_ns[n]
chi_k = np.random.randn(K)
chi_lminus1 = np.random.randn(L)
H_n = get_hurst_value(t_n, list_hurst, ind_hurst)
part1 = 1./(np.random.gamma(0.5 + H_n))
sums1 = np.dot((t_n - t_ks)**(H_n - 0.5) - ((-t_ks)**(H_n - 0.5)), chi_k)
sums2 = np.dot((t_n - t_ns[:n])**(H_n - 0.5), chi_lminus1[:n])
return part1*(1./np.sqrt(L))*(sums1 + sums2)
for n in range(1, L + 1):
onelist.append(multifrac_Brownian_motion(n, L, K, list_hurst, ind_hurst=ind_hurst))
Update:
def list_hurst_funcs(M, seg_size=10):
"""Generate a list of Hurst function components
Args:
M: Int, number of hurst functions
seg_size: Int, number of segmentations of interval [0, 1]
Returns:
list_hurst: List, list of hurst function components
"""
list_hurst = []
for i in range(M):
seg_points = sorted(np.random.uniform(size=seg_size))
funclist = np.random.uniform(size=seg_size + 1)
list_hurst.append((seg_points, funclist))
return list_hurst
def get_hurst_value(x, list_hurst, ind):
if np.isscalar(x):
x = np.array(float(x), ndmin=1)
seg_points, funclist = list_hurst[ind]
condlist = [x < seg_points[0]] +\
[(x >= seg_points[s] and x < seg_points[s + 1])
for s in range(len(seg_points) - 1)] +\
[x >= seg_points[-1]]
return np.piecewise(x, condlist=condlist, funclist=funclist)
One way to tackle a problem like this is to (try) understand the big picture, and come with a different approach that treats everything as 2d or larger (LxK arrays). Another is to examine the multifrac_Brownian_motion, trying to speed it up, and where possible eliminate steps that depend on scalars or 1d arrays. In other words, work from the inside out. If we get enough of a speed improvement it may not matter that we have to call it in a loop. Even better the improvement suggests ways of operating in high dimensions.
As a start from inside out, I'd suggest replacing the t_ks calc with:
t_ks = -np.arange(K,0,-1)/L # 1./L if required by Py2 integer division
Since list_hurst, ind_hurst are the same for all n, I suspect you can move some time consuming parts of get_hurst_value outside the loop.
But I'd put most effort into improving that condlist construction. That's list comprehension buried deep inside your outer loop.
piecewise also loops over those seg_points.
So I'm writing a program in Python to get the GCD of any amount of numbers.
def GCD(numbers):
if numbers[-1] == 0:
return numbers[0]
# i'm stuck here, this is wrong
for i in range(len(numbers)-1):
print GCD([numbers[i+1], numbers[i] % numbers[i+1]])
print GCD(30, 40, 36)
The function takes a list of numbers.
This should print 2. However, I don't understand how to use the the algorithm recursively so it can handle multiple numbers. Can someone explain?
updated, still not working:
def GCD(numbers):
if numbers[-1] == 0:
return numbers[0]
gcd = 0
for i in range(len(numbers)):
gcd = GCD([numbers[i+1], numbers[i] % numbers[i+1]])
gcdtemp = GCD([gcd, numbers[i+2]])
gcd = gcdtemp
return gcd
Ok, solved it
def GCD(a, b):
if b == 0:
return a
else:
return GCD(b, a % b)
and then use reduce, like
reduce(GCD, (30, 40, 36))
Since GCD is associative, GCD(a,b,c,d) is the same as GCD(GCD(GCD(a,b),c),d). In this case, Python's reduce function would be a good candidate for reducing the cases for which len(numbers) > 2 to a simple 2-number comparison. The code would look something like this:
if len(numbers) > 2:
return reduce(lambda x,y: GCD([x,y]), numbers)
Reduce applies the given function to each element in the list, so that something like
gcd = reduce(lambda x,y:GCD([x,y]),[a,b,c,d])
is the same as doing
gcd = GCD(a,b)
gcd = GCD(gcd,c)
gcd = GCD(gcd,d)
Now the only thing left is to code for when len(numbers) <= 2. Passing only two arguments to GCD in reduce ensures that your function recurses at most once (since len(numbers) > 2 only in the original call), which has the additional benefit of never overflowing the stack.
You can use reduce:
>>> from fractions import gcd
>>> reduce(gcd,(30,40,60))
10
which is equivalent to;
>>> lis = (30,40,60,70)
>>> res = gcd(*lis[:2]) #get the gcd of first two numbers
>>> for x in lis[2:]: #now iterate over the list starting from the 3rd element
... res = gcd(res,x)
>>> res
10
help on reduce:
>>> reduce?
Type: builtin_function_or_method
reduce(function, sequence[, initial]) -> value
Apply a function of two arguments cumulatively to the items of a sequence,
from left to right, so as to reduce the sequence to a single value.
For example, reduce(lambda x, y: x+y, [1, 2, 3, 4, 5]) calculates
((((1+2)+3)+4)+5). If initial is present, it is placed before the items
of the sequence in the calculation, and serves as a default when the
sequence is empty.
Python 3.9 introduced multiple arguments version of math.gcd, so you can use:
import math
math.gcd(30, 40, 36)
3.5 <= Python <= 3.8.x:
import functools
import math
functools.reduce(math.gcd, (30, 40, 36))
3 <= Python < 3.5:
import fractions
import functools
functools.reduce(fractions.gcd, (30, 40, 36))
A solution to finding out the LCM of more than two numbers in PYTHON is as follow:
#finding LCM (Least Common Multiple) of a series of numbers
def GCD(a, b):
#Gives greatest common divisor using Euclid's Algorithm.
while b:
a, b = b, a % b
return a
def LCM(a, b):
#gives lowest common multiple of two numbers
return a * b // GCD(a, b)
def LCMM(*args):
#gives LCM of a list of numbers passed as argument
return reduce(LCM, args)
Here I've added +1 in the last argument of range() function because the function itself starts from zero (0) to n-1. Click the hyperlink to know more about range() function :
print ("LCM of numbers (1 to 5) : " + str(LCMM(*range(1, 5+1))))
print ("LCM of numbers (1 to 10) : " + str(LCMM(*range(1, 10+1))))
print (reduce(LCMM,(1,2,3,4,5)))
those who are new to python can read more about reduce() function by the given link.
The GCD operator is commutative and associative. This means that
gcd(a,b,c) = gcd(gcd(a,b),c) = gcd(a,gcd(b,c))
So once you know how to do it for 2 numbers, you can do it for any number
To do it for two numbers, you simply need to implement Euclid's formula, which is simply:
// Ensure a >= b >= 1, flip a and b if necessary
while b > 0
t = a % b
a = b
b = t
end
return a
Define that function as, say euclid(a,b). Then, you can define gcd(nums) as:
if (len(nums) == 1)
return nums[1]
else
return euclid(nums[1], gcd(nums[:2]))
This uses the associative property of gcd() to compute the answer
Try calling the GCD() as follows,
i = 0
temp = numbers[i]
for i in range(len(numbers)-1):
temp = GCD(numbers[i+1], temp)
My way of solving it in Python. Hope it helps.
def find_gcd(arr):
if len(arr) <= 1:
return arr
else:
for i in range(len(arr)-1):
a = arr[i]
b = arr[i+1]
while b:
a, b = b, a%b
arr[i+1] = a
return a
def main(array):
print(find_gcd(array))
main(array=[8, 18, 22, 24]) # 2
main(array=[8, 24]) # 8
main(array=[5]) # [5]
main(array=[]) # []
Some dynamics how I understand it:
ex.[8, 18] -> [18, 8] -> [8, 2] -> [2, 0]
18 = 8x + 2 = (2y)x + 2 = 2z where z = xy + 1
ex.[18, 22] -> [22, 18] -> [18, 4] -> [4, 2] -> [2, 0]
22 = 18w + 4 = (4x+2)w + 4 = ((2y)x + 2)w + 2 = 2z
As of python 3.9 beta 4, it has got built-in support for finding gcd over a list of numbers.
Python 3.9.0b4 (v3.9.0b4:69dec9c8d2, Jul 2 2020, 18:41:53)
[Clang 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import math
>>> A = [30, 40, 36]
>>> print(math.gcd(*A))
2
One of the issues is that many of the calculations only work with numbers greater than 1. I modified the solution found here so that it accepts numbers smaller than 1. Basically, we can re scale the array using the minimum value and then use that to calculate the GCD of numbers smaller than 1.
# GCD of more than two (or array) numbers - alows folating point numbers
# Function implements the Euclidian algorithm to find H.C.F. of two number
def find_gcd(x, y):
while(y):
x, y = y, x % y
return x
# Driver Code
l_org = [60e-6, 20e-6, 30e-6]
min_val = min(l_org)
l = [item/min_val for item in l_org]
num1 = l[0]
num2 = l[1]
gcd = find_gcd(num1, num2)
for i in range(2, len(l)):
gcd = find_gcd(gcd, l[i])
gcd = gcd * min_val
print(gcd)
HERE IS A SIMPLE METHOD TO FIND GCD OF 2 NUMBERS
a = int(input("Enter the value of first number:"))
b = int(input("Enter the value of second number:"))
c,d = a,b
while a!=0:
b,a=a,b%a
print("GCD of ",c,"and",d,"is",b)
As You said you need a program who would take any amount of numbers
and print those numbers' HCF.
In this code you give numbers separated with space and click enter to get GCD
num =list(map(int,input().split())) #TAKES INPUT
def print_factors(x): #MAKES LIST OF LISTS OF COMMON FACTROS OF INPUT
list = [ i for i in range(1, x + 1) if x % i == 0 ]
return list
p = [print_factors(numbers) for numbers in num]
result = set(p[0])
for s in p[1:]: #MAKES THE SET OF COMMON VALUES IN LIST OF LISTS
result.intersection_update(s)
for values in result:
values = values*values #MULTIPLY ALL COMMON FACTORS TO FIND GCD
values = values//(list(result)[-1])
print('HCF',values)
Hope it helped
I'm playing with functional capacities of Python 3 and I tried to implement classical algorithm for calculating Hamming numbers. That's the numbers which have as prime factors only 2, 3 or 5. First Hamming numbers are 2, 3, 4, 5, 6, 8, 10, 12, 15, 16, 18, 20 and so on.
My implementation is the following:
def scale(s, m):
return (x*m for x in s)
def merge(s1, s2):
it1, it2 = iter(s1), iter(s2)
x1, x2 = next(it1), next(it2)
if x1 < x2:
x = x1
it = iter(merge(it1, s2))
elif x1 > x2:
x = x2
it = iter(merge(s1, it2))
else:
x = x1
it = iter(merge(it1, it2))
yield x
while True: yield next(it)
def integers():
n = 0
while True:
n += 1
yield n
m2 = scale(integers(), 2)
m3 = scale(integers(), 3)
m5 = scale(integers(), 5)
m23 = merge(m2, m3)
hamming_numbers = merge(m23, m5)
The problem it that merge seems just doesn't work. Before that I implemented Sieve of Eratosthenes the same way, and it worked perfectly okay:
def sieve(s):
it = iter(s)
x = next(it)
yield x
it = iter(sieve(filter(lambda y: x % y, it)))
while True: yield next(it)
This one uses the same techniques as my merge operation. So I can't see any difference. Do you have any ideas?
(I know that all of these can be implemented other ways, but my goal exactly to understand generators and pure functional capabilities, including recursion, of Python, without using class declarations or special pre-built Python functions.)
UPD: For Will Ness here's my implementation of this algorithms in LISP (Racket actually):
(define (scale str m)
(stream-map (lambda (x) (* x m)) str))
(define (integers-from n)
(stream-cons n
(integers-from (+ n 1))))
(define (merge s1 s2)
(let ((x1 (stream-first s1))
(x2 (stream-first s2)))
(cond ((< x1 x2)
(stream-cons x1 (merge (stream-rest s1) s2)))
((> x1 x2)
(stream-cons x2 (merge s1 (stream-rest s2))))
(else
(stream-cons x1 (merge (stream-rest s1) (stream-rest s2)))))))
(define integers (integers-from 1))
(define hamming-numbers
(stream-cons 1 (merge (scale hamming-numbers 2)
(merge (scale hamming-numbers 3)
(scale hamming-numbers 5)))))
Your algorithm is incorrect. Your m2, m3, m5 should be scaling hamming_numbers, not integers.
The major problem is this: your merge() calls next() for both its arguments unconditionally, so both get advanced one step. So after producing the first number, e.g. 2 for the m23 generator, on the next invocation it sees its 1st argument as 4(,6,8,...) and 2nd as 6(,9,12,...). The 3 is already gone. So it always pulls both its arguments, and always returns the head of the 1st (test entry at http://ideone.com/doeX2Q).
Calling iter() is totally superfluous, it adds nothing here. When I remove it (http://ideone.com/7tk85h), the program works exactly the same and produces exactly the same (wrong) output. Normally iter() serves to create a lazy iterator object, but its arguments here are already such generators.
There's no need to call iter() in your sieve() as well (http://ideone.com/kYh7Di). sieve() already defines a generator, and filter() in Python 3 creates an iterator from a function and an iterable (generators are iterable). See also e.g. Difference between Python's Generators and Iterators .
We can do the merge like this, instead:
def merge(s1, s2):
x1, x2 = next(s1), next(s2)
while True:
if x1 < x2:
yield x1
x1 = next(s1)
elif x1 > x2:
yield x2
x2 = next(s2)
else:
yield x1
x1, x2 = next(s1), next(s2)
Recursion in itself is non-essential in defining the sieve() function too. In fact it only serves to obscure there an enormous deficiency of that code. Any prime it produces will be tested by all the primes below it in value - but only those below its square root are truly needed. We can fix it quite easily in a non-recursive(*) style (http://ideone.com/Qaycpe):
def sieve(s): # call as: sieve( integers_from(2))
x = next(s)
yield x
ps = sieve( integers_from(2)) # independent primes supply
p = next(ps)
q = p*p ; print((p,q))
while True:
x = next(s)
while x<q:
yield x
x = next(s)
# here x == q
s = filter(lambda y,p=p: y % p, s) # filter creation postponed
p = next(ps) # until square of p seen in input
q = p*p
(*)(well, actually, this is also recursive, but in a very different way than before)
This is now much, much, much more efficient (see also: Explain this chunk of haskell code that outputs a stream of primes ).
Recursive or not, is just a syntactic characteristic of a code. The actual run-time structures are the same - the filter() adaptors being hoisted on top of an input stream - either at the appropriate moments, or way too soon (so we'd end up with way too many of them).
I will propose this different approach - using Python heapq (min-heapq) with generator (lazy evaluation) (if you don't insist on keeping the merge() function)
from heapq import heappush, heappop
def hamming_numbers(n):
ans = [1]
last = 0
count = 0
while count < n:
x = heappop(ans)
if x > last:
yield x
last = x
count += 1
heappush(ans, 2* x)
heappush(ans, 3* x)
heappush(ans, 5* x)
>>> n = 20
>>> print(list(hamming_numbers(20)))
[1, 2, 3, 4, 5, 6, 8, 9, 10, 12, 15, 16, 18, 20, 24, 25, 27, 30, 32, 36]
Here's the very dumb way:
def divisorGenerator(n):
for i in xrange(1,n/2+1):
if n%i == 0: yield i
yield n
The result I'd like to get is similar to this one, but I'd like a smarter algorithm (this one it's too much slow and dumb :-)
I can find prime factors and their multiplicity fast enough.
I've an generator that generates factor in this way:
(factor1, multiplicity1)
(factor2, multiplicity2)
(factor3, multiplicity3)
and so on...
i.e. the output of
for i in factorGenerator(100):
print i
is:
(2, 2)
(5, 2)
I don't know how much is this useful for what I want to do (I coded it for other problems), anyway I'd like a smarter way to make
for i in divisorGen(100):
print i
output this:
1
2
4
5
10
20
25
50
100
UPDATE: Many thanks to Greg Hewgill and his "smart way" :)
Calculating all divisors of 100000000 took 0.01s with his way against the 39s that the dumb way took on my machine, very cool :D
UPDATE 2: Stop saying this is a duplicate of this post. Calculating the number of divisor of a given number doesn't need to calculate all the divisors. It's a different problem, if you think it's not then look for "Divisor function" on wikipedia. Read the questions and the answer before posting, if you do not understand what is the topic just don't add not useful and already given answers.
Given your factorGenerator function, here is a divisorGen that should work:
def divisorGen(n):
factors = list(factorGenerator(n))
nfactors = len(factors)
f = [0] * nfactors
while True:
yield reduce(lambda x, y: x*y, [factors[x][0]**f[x] for x in range(nfactors)], 1)
i = 0
while True:
f[i] += 1
if f[i] <= factors[i][1]:
break
f[i] = 0
i += 1
if i >= nfactors:
return
The overall efficiency of this algorithm will depend entirely on the efficiency of the factorGenerator.
To expand on what Shimi has said, you should only be running your loop from 1 to the square root of n. Then to find the pair, do n / i, and this will cover the whole problem space.
As was also noted, this is a NP, or 'difficult' problem. Exhaustive search, the way you are doing it, is about as good as it gets for guaranteed answers. This fact is used by encryption algorithms and the like to help secure them. If someone were to solve this problem, most if not all of our current 'secure' communication would be rendered insecure.
Python code:
import math
def divisorGenerator(n):
large_divisors = []
for i in xrange(1, int(math.sqrt(n) + 1)):
if n % i == 0:
yield i
if i*i != n:
large_divisors.append(n / i)
for divisor in reversed(large_divisors):
yield divisor
print list(divisorGenerator(100))
Which should output a list like:
[1, 2, 4, 5, 10, 20, 25, 50, 100]
I think you can stop at math.sqrt(n) instead of n/2.
I will give you example so that you can understand it easily. Now the sqrt(28) is 5.29 so ceil(5.29) will be 6. So I if I will stop at 6 then I will can get all the divisors. How?
First see the code and then see image:
import math
def divisors(n):
divs = [1]
for i in xrange(2,int(math.sqrt(n))+1):
if n%i == 0:
divs.extend([i,n/i])
divs.extend([n])
return list(set(divs))
Now, See the image below:
Lets say I have already added 1 to my divisors list and I start with i=2 so
So at the end of all the iterations as I have added the quotient and the divisor to my list all the divisors of 28 are populated.
Source: How to determine the divisors of a number
Although there are already many solutions to this, I really have to post this :)
This one is:
readable
short
self contained, copy & paste ready
quick (in cases with a lot of prime factors and divisors, > 10 times faster than the accepted solution)
python3, python2 and pypy compliant
Code:
def divisors(n):
# get factors and their counts
factors = {}
nn = n
i = 2
while i*i <= nn:
while nn % i == 0:
factors[i] = factors.get(i, 0) + 1
nn //= i
i += 1
if nn > 1:
factors[nn] = factors.get(nn, 0) + 1
primes = list(factors.keys())
# generates factors from primes[k:] subset
def generate(k):
if k == len(primes):
yield 1
else:
rest = generate(k+1)
prime = primes[k]
for factor in rest:
prime_to_i = 1
# prime_to_i iterates prime**i values, i being all possible exponents
for _ in range(factors[prime] + 1):
yield factor * prime_to_i
prime_to_i *= prime
# in python3, `yield from generate(0)` would also work
for factor in generate(0):
yield factor
An illustrative Pythonic one-liner:
from itertools import chain
from math import sqrt
def divisors(n):
return set(chain.from_iterable((i,n//i) for i in range(1,int(sqrt(n))+1) if n%i == 0))
But better yet, just use sympy:
from sympy import divisors
I like Greg solution, but I wish it was more python like.
I feel it would be faster and more readable;
so after some time of coding I came out with this.
The first two functions are needed to make the cartesian product of lists.
And can be reused whnever this problem arises.
By the way, I had to program this myself, if anyone knows of a standard solution for this problem, please feel free to contact me.
"Factorgenerator" now returns a dictionary. And then the dictionary is fed into "divisors", who uses it to generate first a list of lists, where each list is the list of the factors of the form p^n with p prime.
Then we make the cartesian product of those lists, and we finally use Greg' solution to generate the divisor.
We sort them, and return them.
I tested it and it seem to be a bit faster than the previous version. I tested it as part of a bigger program, so I can't really say how much is it faster though.
Pietro Speroni (pietrosperoni dot it)
from math import sqrt
##############################################################
### cartesian product of lists ##################################
##############################################################
def appendEs2Sequences(sequences,es):
result=[]
if not sequences:
for e in es:
result.append([e])
else:
for e in es:
result+=[seq+[e] for seq in sequences]
return result
def cartesianproduct(lists):
"""
given a list of lists,
returns all the possible combinations taking one element from each list
The list does not have to be of equal length
"""
return reduce(appendEs2Sequences,lists,[])
##############################################################
### prime factors of a natural ##################################
##############################################################
def primefactors(n):
'''lists prime factors, from greatest to smallest'''
i = 2
while i<=sqrt(n):
if n%i==0:
l = primefactors(n/i)
l.append(i)
return l
i+=1
return [n] # n is prime
##############################################################
### factorization of a natural ##################################
##############################################################
def factorGenerator(n):
p = primefactors(n)
factors={}
for p1 in p:
try:
factors[p1]+=1
except KeyError:
factors[p1]=1
return factors
def divisors(n):
factors = factorGenerator(n)
divisors=[]
listexponents=[map(lambda x:k**x,range(0,factors[k]+1)) for k in factors.keys()]
listfactors=cartesianproduct(listexponents)
for f in listfactors:
divisors.append(reduce(lambda x, y: x*y, f, 1))
divisors.sort()
return divisors
print divisors(60668796879)
P.S.
it is the first time I am posting to stackoverflow.
I am looking forward for any feedback.
Here is a smart and fast way to do it for numbers up to and around 10**16 in pure Python 3.6,
from itertools import compress
def primes(n):
""" Returns a list of primes < n for n > 2 """
sieve = bytearray([True]) * (n//2)
for i in range(3,int(n**0.5)+1,2):
if sieve[i//2]:
sieve[i*i//2::i] = bytearray((n-i*i-1)//(2*i)+1)
return [2,*compress(range(3,n,2), sieve[1:])]
def factorization(n):
""" Returns a list of the prime factorization of n """
pf = []
for p in primeslist:
if p*p > n : break
count = 0
while not n % p:
n //= p
count += 1
if count > 0: pf.append((p, count))
if n > 1: pf.append((n, 1))
return pf
def divisors(n):
""" Returns an unsorted list of the divisors of n """
divs = [1]
for p, e in factorization(n):
divs += [x*p**k for k in range(1,e+1) for x in divs]
return divs
n = 600851475143
primeslist = primes(int(n**0.5)+1)
print(divisors(n))
If your PC has tons of memory, a brute single line can be fast enough with numpy:
N = 10000000; tst = np.arange(1, N); tst[np.mod(N, tst) == 0]
Out:
array([ 1, 2, 4, 5, 8, 10, 16,
20, 25, 32, 40, 50, 64, 80,
100, 125, 128, 160, 200, 250, 320,
400, 500, 625, 640, 800, 1000, 1250,
1600, 2000, 2500, 3125, 3200, 4000, 5000,
6250, 8000, 10000, 12500, 15625, 16000, 20000,
25000, 31250, 40000, 50000, 62500, 78125, 80000,
100000, 125000, 156250, 200000, 250000, 312500, 400000,
500000, 625000, 1000000, 1250000, 2000000, 2500000, 5000000])
Takes less than 1s on my slow PC.
Adapted from CodeReview, here is a variant which works with num=1 !
from itertools import product
import operator
def prod(ls):
return reduce(operator.mul, ls, 1)
def powered(factors, powers):
return prod(f**p for (f,p) in zip(factors, powers))
def divisors(num) :
pf = dict(prime_factors(num))
primes = pf.keys()
#For each prime, possible exponents
exponents = [range(i+1) for i in pf.values()]
return (powered(primes,es) for es in product(*exponents))
Old question, but here is my take:
def divs(n, m):
if m == 1: return [1]
if n % m == 0: return [m] + divs(n, m - 1)
return divs(n, m - 1)
You can proxy with:
def divisorGenerator(n):
for x in reversed(divs(n, n)):
yield x
NOTE: For languages that support, this could be tail recursive.
Assuming that the factors function returns the factors of n (for instance, factors(60) returns the list [2, 2, 3, 5]), here is a function to compute the divisors of n:
function divisors(n)
divs := [1]
for fact in factors(n)
temp := []
for div in divs
if fact * div not in divs
append fact * div to temp
divs := divs + temp
return divs
Here's my solution. It seems to be dumb but works well...and I was trying to find all proper divisors so the loop started from i = 2.
import math as m
def findfac(n):
faclist = [1]
for i in range(2, int(m.sqrt(n) + 2)):
if n%i == 0:
if i not in faclist:
faclist.append(i)
if n/i not in faclist:
faclist.append(n/i)
return facts
If you only care about using list comprehensions and nothing else matters to you!
from itertools import combinations
from functools import reduce
def get_devisors(n):
f = [f for f,e in list(factorGenerator(n)) for i in range(e)]
fc = [x for l in range(len(f)+1) for x in combinations(f, l)]
devisors = [1 if c==() else reduce((lambda x, y: x * y), c) for c in set(fc)]
return sorted(devisors)
My solution via generator function is:
def divisor(num):
for x in range(1, num + 1):
if num % x == 0:
yield x
while True:
yield None
Try to calculate square root a given number and then loop range(1,square_root+1).
number = int(input("Enter a Number: "))
square_root = round(number ** (1.0 / 2))
print(square_root)
divisor_list = []
for i in range(1,square_root+1):
if number % i == 0: # Check if mod return 0 if yes then append i and number/i in the list
divisor_list.append(i)
divisor_list.append(int(number/i))
print(divisor_list)
def divisorGen(n): v = n last = [] for i in range(1, v+1) : if n % i == 0 : last.append(i)
I don´t understand why there are so many complicated solutions to this problem.
Here is my take on it:
def divisors(n):
lis =[1]
s = math.ceil(math.sqrt(n))
for g in range(s,1, -1):
if n % g == 0:
lis.append(g)
lis.append(int(n / g))
return (set(lis))
return [x for x in range(n+1) if n/x==int(n/x)]