Munkres algorithm in Python: calculating maximum profit

Question

I'm using Munkres library from https://github.com/bmc/munkres/ to calculate maximum profit (reversed problem).

What do you think about this coding style?

def get_no_of_vowels(str):
    return sum(str.lower().count(c) for c in "aeiuoy")

def get_no_of_consonants(str):
    return sum(str.lower().count(c) for c in "bcdfghjklmnpqrstvwxz")

def count_letters(str):
    return sum(c.isalpha() for c in str)

def compute_ss(name, item):
    if not count_letters(item)%2:
        ss = get_no_of_vowels(name)*1.5
    else:
        ss = get_no_of_consonants(name)
    if gcd(count_letters(name), count_letters(item)) > 1:
        ss*=1.5
    return ss

if __name__ == "__main__":
    with open(sys.argv[1]) as f:
        for line in f:
            names = line.strip().split(';')[0].split(',')
            items = line.strip().split(';')[1].split(',')
            dic = collections.defaultdict(list)
            profit_matrix = []
            for name in names:
                row = []
                for item in items:
                    row.append(compute_ss(name, item))
                profit_matrix.append(row)
            m = Munkres()
            cost_matrix = make_cost_matrix(profit_matrix, lambda x: 1e10 - x)
            indexes = m.compute(cost_matrix)
            total = 0
            for row, column in indexes:
                value = profit_matrix[row][column]
                # print '(%.2f, %.2f) -> %.2f' % (row, column, value)
                total += value
            print "%.2f"%total

Answer 1

I just want to point out two sources of inefficiency:

1) The counting functions, e.g.:

def get_no_of_vowels(str):
    return sum(str.lower().count(c) for c in "aeiuoy")

This will loop over str 6 times, which is 5 more times than is necessary. A more efficient approach is:

def get_no_of_vowels(str):
    return sum(1 for s in str if s.lower() in "aeiouy" else 0)

We are creating a generator that will convert each character into a 1 or a 0, and then just sum them. That way, we just do one iteration. Also, "y" is not a vowel.

2) Your line-splitting code here:

        names = line.strip().split(';')[0].split(',')
        items = line.strip().split(';')[1].split(',')

You are splitting line twice. That's extra work. Prefer:

names, items = line.strip().split(';')
names = names.split(',')
items = items.split(',')

Hope that helps.