Use comprehension in it_IT address provider to populate cities

[sevdog]

What does this change

This PR uses a single set-comprehension to populate the cities list for it_IT locale.

What was wrong

The previous implementation used a series of nested for-loops to build up a unique list of cities in the it_IT locale using a strange inline at line 10 to avoid using a normal IF statement.

faker/faker/providers/address/it_IT/__init__.py

Lines 6 to 11 in 2069bd3

	def getcities(fulldict):
	cities = []
	for cap in fulldict:
	for c in fulldict[cap]:
	cities.append(c[0]) if c[0] not in cities else cities
	return cities

While profiling the code in which I use this provider I found out it was taking a long time to load.

How this fixes it

Using a set-comprehension and then turning that into a list takes advantages of built-in features of set and the usage of .items() for iterating over a dict provide better performances.

Local benchmark tests with timeit shows the latter is at least 100x faster.

from timeit import timeit
from faker.providers.address.it_IT import Provider

fragment = 'getcities(cap_city_province)'

def original_getcities(fulldict):
    cities = []
    for cap in fulldict:
        for c in fulldict[cap]:
            cities.append(c[0]) if c[0] not in cities else cities
    return cities

def new_getcities(fulldict):
    return list({c[0] for _cap, cities in fulldict.items() for c in cities})

print(timeit(fragment, globals={'cap_city_province': Provider.cap_city_province, 'getcities': original_getcities}))
print(timeit(fragment, globals={'cap_city_province': Provider.cap_city_province, 'getcities': new_getcities}))

[fcurella]

Thank you!