graycode_enum.py

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import random
from tqdm import tqdm
import itertools
data = pd.read_csv("law.csv")

from math import factorial

def gray_code_composition(n):
    length = int(factorial(2*n-1)/(factorial(n-1)*factorial(2*n-1-(n-1))))
    compositions = np.zeros((length, n))

    # first combination
    curr = np.zeros(n)
    curr[0] = n
    compositions[0, :] = curr
    p = 0
    pos=1

    while (curr[n-1] != n): # we are done once n is in last
        if p == 0:
            if np.count_nonzero(curr) > 1:
                b = np.flatnonzero(curr)[1]
            else:
                b=0
            if b == 1:
                if curr[0] == 1:
                    p = 1
            elif (n - curr[0]) % 2 == 0:
                d, i, p = 0, 1, 1
            elif curr[b] % 2 == 1:
                d, i, p = 0, b, b
            else:
                i, d = 0, b
        else:
            if (n - curr[p]) % 2 == 1:
                d, i = p, p-1
                if curr[p] % 2 == 0:
                    i = 0
                p = i
            elif curr[p+1] % 2 == 0:
                i, d = p+1, p
                if curr[p] == 1:
                    p= p+1
            else:
                i, d = p, p+1
        curr[i] +=1
        curr[d] -= 1
        if curr[0] > 0:
            p=0
        compositions[pos] = curr
        pos += 1
    return compositions


n = data.shape[0]
graycodes = gray_code_composition(n)

data = data.values

corrs = []

for graycode in tqdm(graycodes):
    indices = list(itertools.chain(*[[i]*int(g) for i, g in enumerate(graycode)]))
    corr = np.corrcoef(data[indices].T)[0][1]
    corrs.append(corr)


pd.DataFrame(corrs).to_csv("graycode_enumeration_corrcoefs.csv", index=False)