best way for convert binary list to list of chars

问题:

i have a special list like this:

[0,0,0,1,0,1,0,1,1,1,0,1]

I want it map to a char list like:

['+','+','+','-','+','-','+','-','-','-','+','-']

here is some code. it multiplies probabilities to calculate joint probabilities:

def solve_all_asigned(self,joint_s):
    mult_val=1
    for tab in self.tables:
        if tab['decision']:
            continue
        name=tab['name']
        nr=name.replace('.',' ')
        nr_spd=re.split(' ',nr)
        val=''
        check_list=[x in joint_s.keys() for x in nr_spd]
        if False in check_list:
            continue
        val=''.join(map(joint_s.get,nr_spd))
        mult_val=mult_val*tab[val]
    return mult_val

n=22
joint_s2={}
all_combinations=list(itertools.product([0,1],repeat=n))
for binlist in all_combinations:
    for i in range(n):
        joint_s2[nan_set[i]]='+' if binlist[i]=='0' else '-'
    s.append(self.solve_all_asigned(joint_s2))
ss=sum(s)

joint_s:{'AA': 'nan', 'AC': 'nan', 'AB': 'nan', 'AE': 'nan', 'AD': 'nan', 'AF': 'nan', 'A': 'nan', 'C': 'nan', 'B': 'nan', 'E': 'nan', 'D': 'nan', 'G': 'nan', 'F': 'nan', 'I': 'nan', 'H': 'nan', 'K': 'nan', 'J': 'nan', 'M': 'nan', 'L': 'nan', 'O': 'nan', 'N': 'nan', 'Q': 'nan', 'P': '+', 'S': 'nan', 'R': 'nan', 'U': 'nan', 'T': 'nan', 'W': 'nan', 'V': 'nan', 'Y': 'nan', 'X': '+', 'Z': 'nan'}
nan_set:['A', 'C', 'B', 'E', 'D', 'G', 'F', 'I', 'H', 'K', 'J', 'M', 'L', 'O', 'N', 'Q', 'S', 'R', 'U', 'T', 'W', 'V']
tab:{'name': 'A.B', '--': 0.19999999999999996, 'decision': False, '+-': 0.6, '++': 0.4, '-+': 0.8}

how can I do this?

回答1:

To answer your original question, here's an efficient way to map a list of 0 & 1 integers to plus and minus signs.

binlist = [0, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1]
a = ['+-'[u] for u in binlist]
print(a)

output

['+', '+', '+', '-', '+', '-', '+', '-', '-', '-', '+', '-']

---

However, as I said above, this may not actually be necessary for your probability calculations.

Here are some other improvements that can be made to your code.

all_combinations=list(itertools.product([0,1],repeat=n))
for binlist in all_combinations:

wastes a lot of RAM. For n = 22, the all_combinations list contains over 4 million 22 item tuples. It's far more efficient to iterate over the output of product directly:

for binlist in itertools.product([0, 1], repeat=n):

In this line:

joint_s2[nan_set[i]]='+' if binlist[i]=='0' else '-'

You test binlist[i] against the string '0', but the binlist tuples contain the integers 0 and 1, not strings, and Python will not automatically convert numeric strings to integers or vice versa.

Another puzzling section is

nr=name.replace('.',' ')
nr_spd=re.split(' ',nr)

Why use regex? Why not just use the built-in str.split method, eg

nr_spd = name.split('.')

OTOH, if your name fields may be separated by one or more spaces as well as by at most a single dot then you can do

nr_spd = name.replace('.', ' ', 1).split()