Python实现的最近最少使用算法


Posted in Python onJuly 10, 2015

本文实例讲述了Python实现的最近最少使用算法。分享给大家供大家参考。具体如下:

# lrucache.py -- a simple LRU (Least-Recently-Used) cache class 
# Copyright 2004 Evan Prodromou <evan@bad.dynu.ca> 
# Licensed under the Academic Free License 2.1 
# Licensed for ftputil under the revised BSD license 
# with permission by the author, Evan Prodromou. Many 
# thanks, Evan! :-) 
# 
# The original file is available at 
# http://pypi.python.org/pypi/lrucache/0.2 . 
# arch-tag: LRU cache main module 
"""a simple LRU (Least-Recently-Used) cache module 
This module provides very simple LRU (Least-Recently-Used) cache 
functionality. 
An *in-memory cache* is useful for storing the results of an 
'expe\nsive' process (one that takes a lot of time or resources) for 
later re-use. Typical examples are accessing data from the filesystem, 
a database, or a network location. If you know you'll need to re-read 
the data again, it can help to keep it in a cache. 
You *can* use a Python dictionary as a cache for some purposes. 
However, if the results you're caching are large, or you have a lot of 
possible results, this can be impractical memory-wise. 
An *LRU cache*, on the other hand, only keeps _some_ of the results in 
memory, which keeps you from overusing resources. The cache is bounded 
by a maximum size; if you try to add more values to the cache, it will 
automatically discard the values that you haven't read or written to 
in the longest time. In other words, the least-recently-used items are 
discarded. [1]_ 
.. [1]: 'Discarded' here means 'removed from the cache'. 
"""
from __future__ import generators 
import time 
from heapq import heappush, heappop, heapify 
# the suffix after the hyphen denotes modifications by the 
# ftputil project with respect to the original version 
__version__ = "0.2-1"
__all__ = ['CacheKeyError', 'LRUCache', 'DEFAULT_SIZE'] 
__docformat__ = 'reStructuredText en'
DEFAULT_SIZE = 16
"""Default size of a new LRUCache object, if no 'size' argument is given."""
class CacheKeyError(KeyError): 
  """Error raised when cache requests fail 
  When a cache record is accessed which no longer exists (or never did), 
  this error is raised. To avoid it, you may want to check for the existence 
  of a cache record before reading or deleting it."""
  pass
class LRUCache(object): 
  """Least-Recently-Used (LRU) cache. 
  Instances of this class provide a least-recently-used (LRU) cache. They 
  emulate a Python mapping type. You can use an LRU cache more or less like 
  a Python dictionary, with the exception that objects you put into the 
  cache may be discarded before you take them out. 
  Some example usage:: 
  cache = LRUCache(32) # new cache 
  cache['foo'] = get_file_contents('foo') # or whatever 
  if 'foo' in cache: # if it's still in cache... 
    # use cached version 
    contents = cache['foo'] 
  else: 
    # recalculate 
    contents = get_file_contents('foo') 
    # store in cache for next time 
    cache['foo'] = contents 
  print cache.size # Maximum size 
  print len(cache) # 0 <= len(cache) <= cache.size 
  cache.size = 10 # Auto-shrink on size assignment 
  for i in range(50): # note: larger than cache size 
    cache[i] = i 
  if 0 not in cache: print 'Zero was discarded.' 
  if 42 in cache: 
    del cache[42] # Manual deletion 
  for j in cache:  # iterate (in LRU order) 
    print j, cache[j] # iterator produces keys, not values 
  """
  class __Node(object): 
    """Record of a cached value. Not for public consumption."""
    def __init__(self, key, obj, timestamp, sort_key): 
      object.__init__(self) 
      self.key = key 
      self.obj = obj 
      self.atime = timestamp 
      self.mtime = self.atime 
      self._sort_key = sort_key 
    def __cmp__(self, other): 
      return cmp(self._sort_key, other._sort_key) 
    def __repr__(self): 
      return "<%s %s => %s (%s)>" % \ 
          (self.__class__, self.key, self.obj, \ 
          time.asctime(time.localtime(self.atime))) 
  def __init__(self, size=DEFAULT_SIZE): 
    # Check arguments 
    if size <= 0: 
      raise ValueError, size 
    elif type(size) is not type(0): 
      raise TypeError, size 
    object.__init__(self) 
    self.__heap = [] 
    self.__dict = {} 
    """Maximum size of the cache. 
    If more than 'size' elements are added to the cache, 
    the least-recently-used ones will be discarded."""
    self.size = size 
    self.__counter = 0
  def _sort_key(self): 
    """Return a new integer value upon every call. 
    Cache nodes need a monotonically increasing time indicator. 
    time.time() and time.clock() don't guarantee this in a 
    platform-independent way. 
    """
    self.__counter += 1
    return self.__counter 
  def __len__(self): 
    return len(self.__heap) 
  def __contains__(self, key): 
    return self.__dict.has_key(key) 
  def __setitem__(self, key, obj): 
    if self.__dict.has_key(key): 
      node = self.__dict[key] 
      # update node object in-place 
      node.obj = obj 
      node.atime = time.time() 
      node.mtime = node.atime 
      node._sort_key = self._sort_key() 
      heapify(self.__heap) 
    else: 
      # size may have been reset, so we loop 
      while len(self.__heap) >= self.size: 
        lru = heappop(self.__heap) 
        del self.__dict[lru.key] 
      node = self.__Node(key, obj, time.time(), self._sort_key()) 
      self.__dict[key] = node 
      heappush(self.__heap, node) 
  def __getitem__(self, key): 
    if not self.__dict.has_key(key): 
      raise CacheKeyError(key) 
    else: 
      node = self.__dict[key] 
      # update node object in-place 
      node.atime = time.time() 
      node._sort_key = self._sort_key() 
      heapify(self.__heap) 
      return node.obj 
  def __delitem__(self, key): 
    if not self.__dict.has_key(key): 
      raise CacheKeyError(key) 
    else: 
      node = self.__dict[key] 
      del self.__dict[key] 
      self.__heap.remove(node) 
      heapify(self.__heap) 
      return node.obj 
  def __iter__(self): 
    copy = self.__heap[:] 
    while len(copy) > 0: 
      node = heappop(copy) 
      yield node.key 
    raise StopIteration 
  def __setattr__(self, name, value): 
    object.__setattr__(self, name, value) 
    # automagically shrink heap on resize 
    if name == 'size': 
      while len(self.__heap) > value: 
        lru = heappop(self.__heap) 
        del self.__dict[lru.key] 
  def __repr__(self): 
    return "<%s (%d elements)>" % (str(self.__class__), len(self.__heap)) 
  def mtime(self, key): 
    """Return the last modification time for the cache record with key. 
    May be useful for cache instances where the stored values can get 
    'stale', such as caching file or network resource contents."""
    if not self.__dict.has_key(key): 
      raise CacheKeyError(key) 
    else: 
      node = self.__dict[key] 
      return node.mtime 
if __name__ == "__main__": 
  cache = LRUCache(25) 
  print cache 
  for i in range(50): 
    cache[i] = str(i) 
  print cache 
  if 46 in cache: 
    print "46 in cache"
    del cache[46] 
  print cache 
  cache.size = 10
  print cache 
  cache[46] = '46'
  print cache 
  print len(cache) 
  for c in cache: 
    print c 
  print cache 
  print cache.mtime(46) 
  for c in cache: 
    print c

希望本文所述对大家的Python程序设计有所帮助。

Python 相关文章推荐
python 队列详解及实例代码
Oct 18 Python
Python变量和数据类型详解
Feb 15 Python
python中OrderedDict的使用方法详解
May 05 Python
Python实现二维数组按照某行或列排序的方法【numpy lexsort】
Sep 22 Python
python命令行解析之parse_known_args()函数和parse_args()使用区别介绍
Jan 24 Python
Python中的单继承与多继承实例分析
May 10 Python
使用python存储网页上的图片实例
May 22 Python
Python统计python文件中代码,注释及空白对应的行数示例【测试可用】
Jul 25 Python
Python+OpenCV 实现图片无损旋转90°且无黑边
Dec 12 Python
Tensorflow累加的实现案例
Feb 05 Python
Python实现异步IO的示例
Nov 05 Python
python代码实现扫码关注公众号登录的实战
Nov 01 Python
Python导入oracle数据的方法
Jul 10 #Python
Python验证码识别的方法
Jul 10 #Python
Python实现大文件排序的方法
Jul 10 #Python
Python实现telnet服务器的方法
Jul 10 #Python
Python读写unicode文件的方法
Jul 10 #Python
Python实现提取谷歌音乐搜索结果的方法
Jul 10 #Python
python和bash统计CPU利用率的方法
Jul 10 #Python
You might like
不用数据库的多用户文件自由上传投票系统(3)
2006/10/09 PHP
精通php的十大要点(上)
2009/02/04 PHP
php中数据库连接方式pdo和mysqli对比分析
2015/02/25 PHP
php通过asort()给关联数组按照值排序的方法
2015/03/18 PHP
JavaScript call apply使用 JavaScript对象的方法绑定到DOM事件后this指向问题
2011/09/28 Javascript
javascript中数组的冒泡排序使用示例
2013/12/18 Javascript
javascript实现在下拉列表中显示多级树形菜单的方法
2015/08/12 Javascript
jquery实现的动态回到顶部特效代码
2015/10/28 Javascript
Bootstrap编写一个兼容主流浏览器的受众巨幕式风格页面
2016/07/01 Javascript
JavaScript日期对象(Date)基本用法示例
2017/01/18 Javascript
简单快速的实现js计算器功能
2017/08/17 Javascript
基于js中的存储键值对以及注意事项介绍
2018/03/30 Javascript
浅谈让你的代码更简短,更整洁,更易读的ES6小技巧
2018/10/25 Javascript
Layui弹出层 加载 做编辑页面的方法
2019/09/16 Javascript
使用element-ui +Vue 解决 table 里包含表单验证的问题
2020/07/17 Javascript
python写的ARP攻击代码实例
2014/06/04 Python
用Python进行一些简单的自然语言处理的教程
2015/03/31 Python
简单实现python爬虫功能
2015/12/31 Python
Python的Django中将文件上传至七牛云存储的代码分享
2016/06/03 Python
关于Python中Inf与Nan的判断问题详解
2017/02/08 Python
Python操作json的方法实例分析
2018/12/06 Python
Python 实现日志同时输出到屏幕和文件
2020/02/19 Python
pytorch dataloader 取batch_size时候出现bug的解决方式
2020/02/20 Python
Django框架models使用group by详解
2020/03/11 Python
科茨沃尔德家居商店:Scotts of Stow
2018/06/29 全球购物
Anthropologie英国:美国家喻户晓的休闲服装和家居产品品牌
2018/12/05 全球购物
Tom Dixon官网:英国照明及家具设计和制造公司
2019/03/01 全球购物
初始化了一个没有run()方法的线程类,是否会出错?
2014/03/27 面试题
童装店创业计划书
2014/01/09 职场文书
2014机关干部学习“焦裕禄精神”思想汇报
2014/09/19 职场文书
校园环境卫生倡议书
2015/04/29 职场文书
步步惊心观后感
2015/06/12 职场文书
房屋产权证明书
2015/06/19 职场文书
导游词之澳门玫瑰圣母堂
2019/12/03 职场文书
在Java中Collection的一些常用方法总结
2021/06/13 Java/Android
使用ICOM IC-R9500接收机同时测评十台收音机中波接收性能
2022/05/10 无线电