避免Python内存泄漏的几种有效方法

class Node:
    def __init__(self, value):
        self.value = value
        self.ref = None
 
def create_cycle():
    node1 = Node(1)
    node2 = Node(2)
    node1.ref = node2
    node2.ref = node1  # 创建循环引用
 
create_cycle()

import gc
 
class Node:
    def __init__(self, value):
        self.value = value
        self.ref = None
 
def create_cycle():
    node1 = Node(1)
    node2 = Node(2)
    node1.ref = node2
    node2.ref = node1  # 创建循环引用
 
gc.set_debug(gc.DEBUG_LEAK)  # 启用调试模式
create_cycle()
 
gc.collect()  # 强制进行垃圾回收

import objgraph
import gc
 
class Node:
    def __init__(self, value):
        self.value = value
        self.ref = None
 
def create_cycle():
    node1 = Node(1)
    node2 = Node(2)
    node1.ref = node2
    node2.ref = node1  # 创建循环引用
 
create_cycle()
 
gc.collect()  # 强制进行垃圾回收
objgraph.show_backrefs([node1], filename='refs.png')  # 生成对象引用图

pip install memory_profiler

from memory_profiler import profile
 
@profile
def create_large_list():
    large_list = [i for i in range(100000)]
    return large_list
 
create_large_list()

import weakref
 
class Node:
    def __init__(self, value):
        self.value = value
        self.ref = None
 
def create_cycle():
    node1 = Node(1)
    node2 = Node(2)
    node1.ref = weakref.ref(node2)  # 使用弱引用
    node2.ref = weakref.ref(node1)
 
create_cycle()

def read_large_file(file_path):
    with open(file_path, 'r') as file:
        data = file.read()
    return data

class ObjectPool:
    def __init__(self):
        self.pool = []
 
    def get(self):
        if self.pool:
            return self.pool.pop()
        else:
            return self.create_new()
 
    def release(self, obj):
        self.pool.append(obj)
 
    def create_new(self):
        return Node(0)
 
# 使用对象池
pool = ObjectPool()
node = pool.get()
pool.release(node)

import numpy as np
 
# 使用numpy数组代替列表
large_array = np.arange(1000000)