Suggestions for writing this small piece of code more elegantly

Question

Although it looks terrible, I am not finding a nicer/more efficient way of doing this:

ae    = np.arange(0.0,1,0.05)
aee   = np.arange(0.3,1.01,0.345)
aef   = np.arange(0.3,1.01,0.345)
random.shuffle(ae)
random.shuffle(aee)
random.shuffle(aef)
for item_a in aee:
    for item_b in ae:
        for item_c in aef: 
            rlist.append(colorsys.hsv_to_rgb(item_b,item_a,item_c))

Ideas?

Answer 1

If you do not want to shuffle the rlist, but the initial lists, then you can try to put the last four lines into a list comprehension:

rlist = [ colorsys.hsv_to_rgb(b, a, c) for c in aef for b in ae for a in aee ] 

Answer 2

import numpy as np
import random
import itertools
import colorsys
hue, saturation, value = np.arange(0.0,1,0.05), np.arange(0.3,1.01,0.345), np.arange(0.3,1.01,0.345)
rlist= [colorsys.hsv_to_rgb(hue, saturation, value) for hue, saturation, value in
        itertools.product(random.sample(hue,len(hue)), random.sample(saturation, len(saturation)), random.sample(value, len(value)))]
print rlist

EDIT: random.sample from full population to avoid inplace separate shuffles

The version without itertools:

# without itertools
import numpy as np
import random
from pprint import pprint
import colorsys
hues, saturations, values = np.arange(0.0,1,0.05), np.arange(0.3,1.01,0.345), np.arange(0.3,1.01,0.345)
rlist= [colorsys.hsv_to_rgb(hue, saturation, value)
        for hue in random.sample(hues,len(hues))
        for saturation in random.sample(saturations, len(saturations))
        for value in random.sample(values, len(values))]
pprint(rlist)

You can also include the definition of itertools.product from documentation (I did that in module called it.py in my server and used it instead of itertools):

product = None
from itertools import *
if not product:
    def product(*args, **kwds):
        # product('ABCD', 'xy') --> Ax Ay Bx By Cx Cy Dx Dy
        # product(range(2), repeat=3) --> 000 001 010 011 100 101 110 111
        pools = map(tuple, args) * kwds.get('repeat', 1)
        result = [[]]
        for pool in pools:
            result = [x+[y] for x in result for y in pool]
        for prod in result:
            yield tuple(prod)

I use itertools normally as:

import itertools as it

But in the server it is replaced by

import it

Answer 3

you don't need to shuffle each list in the first place because you will do a Cartesian product ... ;

import itertools
import colorsys


hsv_iter = itertools.product(np.arange((0, 1, 0.05),
                             np.arange((0.3,1.01,0.345),
                             np.arange((0.3,1.01,0.345))

rlist = [colorsys.hsv_to_rgb(hue, lightness, saturation)
         for hue, lightness, saturation in hsv_ite]

# you can shuffle now the list if you want
random.shuffle(rlist)

Answer 4

Stupid oneliner:

rlist = [colorsys.hsv_to_rgb(b, a, c) for c in random.sample(aef,len(aef))
                                      for b in random.sample(ae,len(ae)) 
                                      for a in random.sample(aee,len(aee))] 

random.sample(x,len(x)) (new in 2.3) is pretty much equivalent to random.shuffle(x), but it returns a randomized copy of the list instead of None.

This is probably much slower than shuffling or whatever else, and you don't keep a copy of the randomized lists (if you care).