我可以在python中有一个全局变量而不将其声明为全局变量吗？

Question

I want some sort of a global state that maintains the current number, as well as a function to generate the next number.

I can write a generator to give me the next number.

 def gen(self, n):
    yield n
    yield n + 1

but what is a clean way to maintain its state? I do not want to simply have a global variable. Is there a better way to do this? or is that my only option?

I tried to make a class like this:

class Count:
"""
Represents the counter which
generates variables
"""

def __init__(self, curr=0):
    """
    :param curr: the current integer
    """
    self.curr = curr

def gen(self):
    """
    A generator for the next
    number
    :return: generator
    """
    self.curr += 1
    yield self.curr
    yield self.curr + 1

but this will not work, because every time I create Count(), it will reset my counter which I don't want.

Answer 1

If my understanding is correct, to eliminate the global counter you could create a closure for you variable and return a function that increments it.

The original function counter is called only once, consecutive calls simply increment the counter:

def count(n):
    c = n
    def incr():
        nonlocal c
        c += 1
        print(c)
        return c
    return incr

count is initialized with some state n and incr references that state in consecutive calls:

>>> f = count(2)
>>> f()
3
>>> f()
4
>>> f()
5

Answer 2

Or you can use the next(generator) function.

def num_gen():
    n=1
    while n:
        yield n
        n += 1

Then

>>my_gen = num_gen()
>>next(my_gen)
1
>>next(my_gen)
2

And so forth. Whenever a generator yields a value, the state of the generator is stored so that it can resume execution later.

Answer 3

If you want to maintain state across multiple instances of Count , then use a variable in the class scope, and reference it with the Count. prefix, like this:

class Count:
    curr = 0

    def __init__(self, startWith = None):
        if startWith is not None: Count.curr = startWith - 1

    def gen(self):
        while True:
            Count.curr += 1
            yield Count.curr

Note that if you want to maintain state, the constructor should allow for the possibility to not reset the counter, but leave it untouched.

As a side note, you might be interested in letting the generator generate a never ending series as shown above.

Here is how you could use the above class:

# Get generator that starts with 2
gen = Count(2).gen();

print (next(gen)); # -> 2
print (next(gen)); # -> 3
print (next(gen)); # -> 4

# Get a generator that continues ...
gen2 = Count().gen();

print (next(gen2)); # -> 5

# We can still use the previous generator also:

print (next(gen)); # -> 6

# Start with 0:

gen3 = Count(0).gen();

print (next(gen3)); # -> 0

# Other generators follow suit:

print (next(gen)); # -> 1
print (next(gen2)); # -> 2