导入的 python 函数在从模块 1 导入时有效，但在从模块 2 导入时失败

Question

我有一个奇怪的问题想了解。我写了一个简单的遗传密码来猜测这对字母“ab”。有一个循环使用生成随机候选的函数作为第一步。该函数可以工作并且很容易理解，但是，它在导入时不起作用，只有在它位于整个遗传密码模块内部的情况下才起作用。如果我编写另一个具有该功能的模块，然后从那里导入，它就可以正常工作。

这到底是什么？

这是我用来导入函数的代码：

from scratch_2  import generate_random_candidate

print (generate_random_candidate())

这是scratch_2：

import random
import string

target = "ab"
letras = string.ascii_lowercase
tried_candidates = set()

def generate_random_candidate():
    while True:
        intento = random.choice(letras) + random.choice(letras)
        if intento not in tried_candidates:
            tried_candidates.add(intento)
            return intento

所以这就是完全正常的方法。这是scratch_1，当我从那里导入时，我看到调试功能不起作用，如果我打印tried_candidates，我看到它打印set()、set()、set()，就像它在无限循环内一样。是因为它调用了该函数但没有运行其代码吗？与变量tried_candidates有关吗？相同的功能不应该独立于存储在其中的模块而工作吗？

这是scratch_1，如果我从中导入函数，它什么也不做。

import random
import string
import matplotlib.pyplot as plt
import plotly.graph_objects as go
target = "ab"
letras = string.ascii_lowercase
tried_candidates = set()


def generate_random_candidate():
    while True:
        intento = random.choice(letras) + random.choice(letras)
        if intento not in tried_candidates:
            tried_candidates.add(intento)
            return intento


def generate_children(parent):
    hijos = []
    while True:
        intento = parent[0] + random.choice(letras)
        if intento not in tried_candidates:
            hijos.append(intento)
            break
    while True:
        intento = random.choice(letras) + parent[1]
        if intento not in tried_candidates:
            hijos.append(intento)
            break
    return hijos

def evaluate_candidate(candidate):
    score = 0
    if candidate[0] == target[0]:
        score += 1
    if candidate[1] == target[1]:
        score += 1
    """
    if candidate[0] == target[1]:
        score += 0.1
    if candidate[1] == target[0]:
        score += 0.1
    """

    return score

def g ():
    tried_candidates = set()
    population_size = 10
    population = []

    for _ in range(population_size):
        population.append(generate_random_candidate())
    contador = 1

    while True:
        tried_candidates.update(population)
        scores = []
        for candidate in population:
            scores.append(evaluate_candidate(candidate))
        children = []
        Flag = False
        for i, score in enumerate(scores):
            if score == 2:
                Flag = True
                break
            if score == 1 :
                x = generate_children(population[i])
                tried_candidates.update(x)
                children.extend(generate_children(population[i]))

        if len(children) < population_size:
            new_candidates = []
            for _ in range(population_size - len(children)):
                y = generate_random_candidate()
                tried_candidates.update(y)
                new_candidates.append(y)

            children.extend(new_candidates)
            population = children

        else:
            meter = children[population_size:]
            population = children[:population_size]
            tried_candidates.difference_update(meter)

        found = False
        for candidate in population:
            if evaluate_candidate(candidate) == 2 or Flag == True:
                found = True
                return contador

        contador = contador +1

resultados = []
for i in range (100000):
    resultados.append(g())

Answer 1

这里的问题与

tried_candidates

变量有关。由于

tried_candidates

是在

scratch_1

和

scratch_2

中在模块级别定义的一组，因此将

generate_random_candidate

从一个模块导入另一个模块可能会导致不同的实例或行为，尤其是在每次调用中修改

set

时.

这是修改后的代码：

# scratch_1.py
import random
import string
import matplotlib.pyplot as plt
import plotly.graph_objects as go

target = "ab"
letras = string.ascii_lowercase

def generate_random_candidate(tried_candidates):
    while True:
        intento = random.choice(letras) + random.choice(letras)
        if intento not in tried_candidates:
            tried_candidates.add(intento)
            return intento

def generate_children(parent, tried_candidates):
    hijos = []
    while True:
        intento = parent[0] + random.choice(letras)
        if intento not in tried_candidates:
            hijos.append(intento)
            break
    while True:
        intento = random.choice(letras) + parent[1]
        if intento not in tried_candidates:
            hijos.append(intento)
            break
    return hijos

def g():
    tried_candidates = set()
    population_size = 10
    population = []

    for _ in range(population_size):
        population.append(generate_random_candidate(tried_candidates))
    contador = 1

    while True:
        scores = []
        for candidate in population:
            scores.append(evaluate_candidate(candidate))
        children = []
        Flag = False
        for i, score in enumerate(scores):
            if score == 2:
                Flag = True
                break
            if score == 1:
                x = generate_children(population[i], tried_candidates)
                tried_candidates.update(x)
                children.extend(x)

        if len(children) < population_size:
            new_candidates = []
            for _ in range(population_size - len(children)):
                y = generate_random_candidate(tried_candidates)
                tried_candidates.add(y)
                new_candidates.append(y)

            children.extend(new_candidates)
            population = children

        else:
            meter = children[population_size:]
            population = children[:population_size]
            tried_candidates.difference_update(meter)

        found = False
        for candidate in population:
            if evaluate_candidate(candidate) == 2 or Flag:
                found = True
                return contador

        contador += 1

def evaluate_candidate(candidate):
    score = 0
    if candidate[0] == target[0]:
        score += 1
    if candidate[1] == target[1]:
        score += 1
    return score

resultados = []
for i in range(100000):
    resultados.append(g())

希望这对你有一点帮助。