Python 3之潜藏的利器

bool:"],[20,"\n","36:0"],[20," return \"animal\" in sentence"],[20,"\n","36:0"],[20,"sentence_has_animal(\"Donald had a farm without animals\")"],[20,"\n","36:0"],[20,"# True"],[20,"\n","36:0"],[20,"枚举 (3.4+)"],[20,"\n","32:1"],[20," Python 3 中的 Enum 类支持"],[20,"枚举","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Fenum.html\""],[20,"功能，可以使我们的程序变得更加简洁。 Enum 是一种便捷的变量列表的打包方式，使用该方法能够避免多个变量在代码各处分布，使代码显得杂乱无章。\n\nfrom enum import Enum, auto"],[20,"\n\n","36:0"],[20,"class Monster(Enum):"],[20,"\n","36:0"],[20," ZOMBIE = auto()"],[20,"\n","36:0"],[20," WARRIOR = auto()"],[20,"\n","36:0"],[20," BEAR = auto()"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"print(Monster.ZOMBIE)"],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"\n枚举是一个符号集合，每个符号都和唯一的变量对应。通过使用枚举，我们可以通过符号标识来比较各个成员，我们还可以对枚举本身进行迭代。"],[20,"\n\n","blockquote:true"],[20,"https://docs.python.org/3/library/enum.html"],[20,"\n","blockquote:true"],[20,"\nfor monster in Monster:"],[20,"\n","36:0"],[20," print(monster)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"# Monster.WARRIOR"],[20,"\n","36:0"],[20,"# Monster.BEAR"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"内置的 LRU 缓存（3.2+）"],[20,"\n","32:1"],[20,"在现如今使用的硬件和软件中，缓存无所不在。Python 3 通过 "],[20,"lru_cache","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Ffunctools.html%23functools.lru_cache\""],[20," 来使用 LRU ("],[20,"Least Recently Used","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FCache_replacement_policies%23Least_recently_used_(LRU)\""],[20,") 缓存。\n\n下面的代码定义了一个斐波拉契函数，由于该函数的运算需要多次递归，每次递归都会执行相同的工作，因此使用缓存能够加速它的计算。\n\nimport time"],[20,"\n\n","36:0"],[20,"def fib(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib(number-1) + fib(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 30.684099674224854s"],[20,"\n","36:0"],[20,"\n我们可以使用 lru_cache 来优化该运算。这种优化极技术称为 "],[20,"memoization","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FMemoization\""],[20," ，它能够把执行时间从几秒缩减到几纳秒。\n\nfrom functools import lru_cache"],[20,"\n\n","36:0"],[20,"@lru_cache(maxsize=512)"],[20,"\n","36:0"],[20,"def fib_memoization(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib_memoization(number-1) + fib_memoization(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib_memoization(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 6.866455078125e-05s"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"扩展的可迭代解析功能（3.0+）"],[20,"\n","32:1"],[20,"这里不做详细解释，直接贴出示例代码，具体解决可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-3132%2F\""],[20,"。\nhead, *body, tail = range(5)"],[20,"\n","36:0"],[20,"print(head, body, tail)"],[20,"\n","36:0"],[20,"# 0 [1, 2, 3] 4"],[20,"\n\n","36:0"],[20,"py, filename, *cmds = \"python3.7 script.py -n 5 -l 15\".split()"],[20,"\n","36:0"],[20,"print(py)"],[20,"\n","36:0"],[20,"print(filename)"],[20,"\n","36:0"],[20,"print(cmds)"],[20,"\n","36:0"],[20,"# python3.7"],[20,"\n","36:0"],[20,"# script.py"],[20,"\n","36:0"],[20,"# ['-n', '5', '-l', '15']"],[20,"\n\n","36:0"],[20,"first, _, third, *_ = range(10)"],[20,"\n","36:0"],[20,"print(first, third)"],[20,"\n","36:0"],[20,"# 0 2"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"数据类（3.7+）"],[20,"\n","32:1"],[20,"Python 3 引入了数据类（data class）。其装饰器会自动生成特征方法，例如 __init__() 和 __repr()__，这能够帮助减少样本代码的数量。在官方"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0557%2F\""],[20,"中，它们被称作 “具有默认值的可变命名元组” 。\n\nclass Armor:"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def __init__(self, armor: float, description: str, level: int = 1):"],[20,"\n","36:0"],[20," self.armor = armor"],[20,"\n","36:0"],[20," self.level = level"],[20,"\n","36:0"],[20," self.description = description"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"#"],[20,"\n","36:0"],[20,"\n接来下我们使用数据类来实现上面的代码：\n\nfrom dataclasses import dataclass"],[20,"\n\n","36:0"],[20,"@dataclass"],[20,"\n","36:0"],[20,"class Armor:"],[20,"\n","36:0"],[20," armor: float"],[20,"\n","36:0"],[20," description: str"],[20,"\n","36:0"],[20," level: int = 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"# Armor(armor=5.2, description='Common armor.', level=2)"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"隐式命名空间包（3.3+）"],[20,"\n","32:1"],[20,"构建Python代码有很多方法，其中一种就是在包（packages）中进行构建（即包含一个 __init__.py 文件的文件夹）。以下示例由 Python 的官方文档提供：\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n"],[20,"在 Python 2 中，上面的每个文件夹都必须有一个 ","0:\"%23333333\""],[20," __init__.py ","31:1"],[20,"文件，该文件用于将其所在文件夹转化为 Python 包。然而在 Python 3 中，通过使用","0:\"%23333333\""],[20,"隐式命名空间包","0:\"%23333333\"|16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F\""],[20,"（Implicit Namespace Package）"],[20,"，这些文件就不再需要了。","0:\"%23333333\""],[20,"\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n注：官方文档"],[20,"PEP 420 Specification","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F%23specification\""],[20," 指出，对于一些常规的包，__init__.py 仍然是需要的，如果将该文件删除，就会把该 Python 包变成一个本地的命名空间包，这将会产生一些额外的限制，具体可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fpackaging.python.org%2Fguides%2Fpackaging-namespace-packages%2F%23native-namespace-packages\""],[20,"。\n总结"],[20,"\n","32:1"],[20,"Python3 中的新功能远远不止如此，大家可以进一步进行探索。本文列出的内容只是一些实用的功能，希望能够对大家有所帮助。\n\n文中的代码链接："],[20,"\n","blockquote:true"],[20,"https://github.com/Weenkus/DataWhatNow-Codes/blob/master/things_you_are_probably_not_using_in_python_3_but_should/python%203%20examples.ipynb"],[20,"\n","blockquote:true"]]">

作者 | Vinko Kodoman
翻译 | Monanfei
编辑 | 阿司匹林，Rachel

【导读】在 python 3 推出后，人们开始逐步将基于Python 2 的代码迁移至 Python 3 。但在迁移过程中，很多代码都未能使用到 Python 3 提供的新功能。本文作者介绍了相关功能的介绍，包括字符串格式化处理、文件路径处理、类型提示、内置 LRU 缓存等等，帮助大家更好地利用 python 3 书写代码。

由于 Python 2 即将退出历史的舞台，许多人都开始将 Python 2 的代码转换为 Python 3 ，但在这一修改过程中，人们似乎只是多加了几个括号，大多数人并没有注意到 Python 3 中激动人心的新功能。本文将介绍 Python 3 中一些有趣的功能，希望这些功能能够帮助大家更加轻松的解决一些问题。

注：文中的代码示例基于 Python 3.7 编写，为方便使用，在每个功能后面都列出了该功能所需的最低 Python 版本。

f-strings (3.6+)

复制代码

bool:"],[20,"\n","36:0"],[20," return \"animal\" in sentence"],[20,"\n","36:0"],[20,"sentence_has_animal(\"Donald had a farm without animals\")"],[20,"\n","36:0"],[20,"# True"],[20,"\n","36:0"],[20,"枚举 (3.4+)"],[20,"\n","32:1"],[20," Python 3 中的 Enum 类支持"],[20,"枚举","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Fenum.html\""],[20,"功能，可以使我们的程序变得更加简洁。 Enum 是一种便捷的变量列表的打包方式，使用该方法能够避免多个变量在代码各处分布，使代码显得杂乱无章。\n\nfrom enum import Enum, auto"],[20,"\n\n","36:0"],[20,"class Monster(Enum):"],[20,"\n","36:0"],[20," ZOMBIE = auto()"],[20,"\n","36:0"],[20," WARRIOR = auto()"],[20,"\n","36:0"],[20," BEAR = auto()"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"print(Monster.ZOMBIE)"],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"\n枚举是一个符号集合，每个符号都和唯一的变量对应。通过使用枚举，我们可以通过符号标识来比较各个成员，我们还可以对枚举本身进行迭代。"],[20,"\n\n","blockquote:true"],[20,"https://docs.python.org/3/library/enum.html"],[20,"\n","blockquote:true"],[20,"\nfor monster in Monster:"],[20,"\n","36:0"],[20," print(monster)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"# Monster.WARRIOR"],[20,"\n","36:0"],[20,"# Monster.BEAR"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"内置的 LRU 缓存（3.2+）"],[20,"\n","32:1"],[20,"在现如今使用的硬件和软件中，缓存无所不在。Python 3 通过 "],[20,"lru_cache","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Ffunctools.html%23functools.lru_cache\""],[20," 来使用 LRU ("],[20,"Least Recently Used","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FCache_replacement_policies%23Least_recently_used_(LRU)\""],[20,") 缓存。\n\n下面的代码定义了一个斐波拉契函数，由于该函数的运算需要多次递归，每次递归都会执行相同的工作，因此使用缓存能够加速它的计算。\n\nimport time"],[20,"\n\n","36:0"],[20,"def fib(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib(number-1) + fib(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 30.684099674224854s"],[20,"\n","36:0"],[20,"\n我们可以使用 lru_cache 来优化该运算。这种优化极技术称为 "],[20,"memoization","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FMemoization\""],[20," ，它能够把执行时间从几秒缩减到几纳秒。\n\nfrom functools import lru_cache"],[20,"\n\n","36:0"],[20,"@lru_cache(maxsize=512)"],[20,"\n","36:0"],[20,"def fib_memoization(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib_memoization(number-1) + fib_memoization(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib_memoization(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 6.866455078125e-05s"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"扩展的可迭代解析功能（3.0+）"],[20,"\n","32:1"],[20,"这里不做详细解释，直接贴出示例代码，具体解决可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-3132%2F\""],[20,"。\nhead, *body, tail = range(5)"],[20,"\n","36:0"],[20,"print(head, body, tail)"],[20,"\n","36:0"],[20,"# 0 [1, 2, 3] 4"],[20,"\n\n","36:0"],[20,"py, filename, *cmds = \"python3.7 script.py -n 5 -l 15\".split()"],[20,"\n","36:0"],[20,"print(py)"],[20,"\n","36:0"],[20,"print(filename)"],[20,"\n","36:0"],[20,"print(cmds)"],[20,"\n","36:0"],[20,"# python3.7"],[20,"\n","36:0"],[20,"# script.py"],[20,"\n","36:0"],[20,"# ['-n', '5', '-l', '15']"],[20,"\n\n","36:0"],[20,"first, _, third, *_ = range(10)"],[20,"\n","36:0"],[20,"print(first, third)"],[20,"\n","36:0"],[20,"# 0 2"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"数据类（3.7+）"],[20,"\n","32:1"],[20,"Python 3 引入了数据类（data class）。其装饰器会自动生成特征方法，例如 __init__() 和 __repr()__，这能够帮助减少样本代码的数量。在官方"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0557%2F\""],[20,"中，它们被称作 “具有默认值的可变命名元组” 。\n\nclass Armor:"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def __init__(self, armor: float, description: str, level: int = 1):"],[20,"\n","36:0"],[20," self.armor = armor"],[20,"\n","36:0"],[20," self.level = level"],[20,"\n","36:0"],[20," self.description = description"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"#"],[20,"\n","36:0"],[20,"\n接来下我们使用数据类来实现上面的代码：\n\nfrom dataclasses import dataclass"],[20,"\n\n","36:0"],[20,"@dataclass"],[20,"\n","36:0"],[20,"class Armor:"],[20,"\n","36:0"],[20," armor: float"],[20,"\n","36:0"],[20," description: str"],[20,"\n","36:0"],[20," level: int = 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"# Armor(armor=5.2, description='Common armor.', level=2)"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"隐式命名空间包（3.3+）"],[20,"\n","32:1"],[20,"构建Python代码有很多方法，其中一种就是在包（packages）中进行构建（即包含一个 __init__.py 文件的文件夹）。以下示例由 Python 的官方文档提供：\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n"],[20,"在 Python 2 中，上面的每个文件夹都必须有一个 ","0:\"%23333333\""],[20," __init__.py ","31:1"],[20,"文件，该文件用于将其所在文件夹转化为 Python 包。然而在 Python 3 中，通过使用","0:\"%23333333\""],[20,"隐式命名空间包","0:\"%23333333\"|16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F\""],[20,"（Implicit Namespace Package）"],[20,"，这些文件就不再需要了。","0:\"%23333333\""],[20,"\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n注：官方文档"],[20,"PEP 420 Specification","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F%23specification\""],[20," 指出，对于一些常规的包，__init__.py 仍然是需要的，如果将该文件删除，就会把该 Python 包变成一个本地的命名空间包，这将会产生一些额外的限制，具体可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fpackaging.python.org%2Fguides%2Fpackaging-namespace-packages%2F%23native-namespace-packages\""],[20,"。\n总结"],[20,"\n","32:1"],[20,"Python3 中的新功能远远不止如此，大家可以进一步进行探索。本文列出的内容只是一些实用的功能，希望能够对大家有所帮助。\n\n文中的代码链接："],[20,"\n","blockquote:true"],[20,"https://github.com/Weenkus/DataWhatNow-Codes/blob/master/things_you_are_probably_not_using_in_python_3_but_should/python%203%20examples.ipynb"],[20,"\n","blockquote:true"]]" style="font-size: 15px; color: rgb(63, 63, 63); letter-spacing: 1px;">对任何一种编程语言来说，字符串处理是一项很重要的内容，字符串处理往往是很多程序的基础部分。由于人工处理字符串非常繁琐，我们更希望用一种结构化的方法来处理它们。在 Python 中，我们一般使用 format 来进行结构化字符串处理，如下所示：
bool:"],[20,"\n","36:0"],[20," return \"animal\" in sentence"],[20,"\n","36:0"],[20,"sentence_has_animal(\"Donald had a farm without animals\")"],[20,"\n","36:0"],[20,"# True"],[20,"\n","36:0"],[20,"枚举 (3.4+)"],[20,"\n","32:1"],[20," Python 3 中的 Enum 类支持"],[20,"枚举","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Fenum.html\""],[20,"功能，可以使我们的程序变得更加简洁。 Enum 是一种便捷的变量列表的打包方式，使用该方法能够避免多个变量在代码各处分布，使代码显得杂乱无章。\n\nfrom enum import Enum, auto"],[20,"\n\n","36:0"],[20,"class Monster(Enum):"],[20,"\n","36:0"],[20," ZOMBIE = auto()"],[20,"\n","36:0"],[20," WARRIOR = auto()"],[20,"\n","36:0"],[20," BEAR = auto()"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"print(Monster.ZOMBIE)"],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"\n枚举是一个符号集合，每个符号都和唯一的变量对应。通过使用枚举，我们可以通过符号标识来比较各个成员，我们还可以对枚举本身进行迭代。"],[20,"\n\n","blockquote:true"],[20,"https://docs.python.org/3/library/enum.html"],[20,"\n","blockquote:true"],[20,"\nfor monster in Monster:"],[20,"\n","36:0"],[20," print(monster)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"# Monster.ZOMBIE"],[20,"\n","36:0"],[20,"# Monster.WARRIOR"],[20,"\n","36:0"],[20,"# Monster.BEAR"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"内置的 LRU 缓存（3.2+）"],[20,"\n","32:1"],[20,"在现如今使用的硬件和软件中，缓存无所不在。Python 3 通过 "],[20,"lru_cache","16:\"https%3A%2F%2Fdocs.python.org%2F3%2Flibrary%2Ffunctools.html%23functools.lru_cache\""],[20," 来使用 LRU ("],[20,"Least Recently Used","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FCache_replacement_policies%23Least_recently_used_(LRU)\""],[20,") 缓存。\n\n下面的代码定义了一个斐波拉契函数，由于该函数的运算需要多次递归，每次递归都会执行相同的工作，因此使用缓存能够加速它的计算。\n\nimport time"],[20,"\n\n","36:0"],[20,"def fib(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib(number-1) + fib(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 30.684099674224854s"],[20,"\n","36:0"],[20,"\n我们可以使用 lru_cache 来优化该运算。这种优化极技术称为 "],[20,"memoization","16:\"https%3A%2F%2Fen.wikipedia.org%2Fwiki%2FMemoization\""],[20," ，它能够把执行时间从几秒缩减到几纳秒。\n\nfrom functools import lru_cache"],[20,"\n\n","36:0"],[20,"@lru_cache(maxsize=512)"],[20,"\n","36:0"],[20,"def fib_memoization(number: int) -> int:"],[20,"\n","36:0"],[20," if number == 0: return 0"],[20,"\n","36:0"],[20," if number == 1: return 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," return fib_memoization(number-1) + fib_memoization(number-2)"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"start = time.time()"],[20,"\n","36:0"],[20,"fib_memoization(40)"],[20,"\n","36:0"],[20,"print(f'Duration: {time.time() - start}s')"],[20,"\n","36:0"],[20,"# Duration: 6.866455078125e-05s"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"扩展的可迭代解析功能（3.0+）"],[20,"\n","32:1"],[20,"这里不做详细解释，直接贴出示例代码，具体解决可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-3132%2F\""],[20,"。\nhead, *body, tail = range(5)"],[20,"\n","36:0"],[20,"print(head, body, tail)"],[20,"\n","36:0"],[20,"# 0 [1, 2, 3] 4"],[20,"\n\n","36:0"],[20,"py, filename, *cmds = \"python3.7 script.py -n 5 -l 15\".split()"],[20,"\n","36:0"],[20,"print(py)"],[20,"\n","36:0"],[20,"print(filename)"],[20,"\n","36:0"],[20,"print(cmds)"],[20,"\n","36:0"],[20,"# python3.7"],[20,"\n","36:0"],[20,"# script.py"],[20,"\n","36:0"],[20,"# ['-n', '5', '-l', '15']"],[20,"\n\n","36:0"],[20,"first, _, third, *_ = range(10)"],[20,"\n","36:0"],[20,"print(first, third)"],[20,"\n","36:0"],[20,"# 0 2"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"数据类（3.7+）"],[20,"\n","32:1"],[20,"Python 3 引入了数据类（data class）。其装饰器会自动生成特征方法，例如 __init__() 和 __repr()__，这能够帮助减少样本代码的数量。在官方"],[20,"文档","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0557%2F\""],[20,"中，它们被称作 “具有默认值的可变命名元组” 。\n\nclass Armor:"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def __init__(self, armor: float, description: str, level: int = 1):"],[20,"\n","36:0"],[20," self.armor = armor"],[20,"\n","36:0"],[20," self.level = level"],[20,"\n","36:0"],[20," self.description = description"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"#"],[20,"\n","36:0"],[20,"\n接来下我们使用数据类来实现上面的代码：\n\nfrom dataclasses import dataclass"],[20,"\n\n","36:0"],[20,"@dataclass"],[20,"\n","36:0"],[20,"class Armor:"],[20,"\n","36:0"],[20," armor: float"],[20,"\n","36:0"],[20," description: str"],[20,"\n","36:0"],[20," level: int = 1"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20," def power(self) -> float:"],[20,"\n","36:0"],[20," return self.armor * self.level"],[20,"\n","36:0"],[20," "],[20,"\n","36:0"],[20,"armor = Armor(5.2, \"Common armor.\", 2)"],[20,"\n","36:0"],[20,"armor.power()"],[20,"\n","36:0"],[20,"# 10.4"],[20,"\n\n","36:0"],[20,"print(armor)"],[20,"\n","36:0"],[20,"# Armor(armor=5.2, description='Common armor.', level=2)"],[20,"\n","36:0"],[20,"\n","32:1"],[20,"隐式命名空间包（3.3+）"],[20,"\n","32:1"],[20,"构建Python代码有很多方法，其中一种就是在包（packages）中进行构建（即包含一个 __init__.py 文件的文件夹）。以下示例由 Python 的官方文档提供：\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," __init__.py"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n"],[20,"在 Python 2 中，上面的每个文件夹都必须有一个 ","0:\"%23333333\""],[20," __init__.py ","31:1"],[20,"文件，该文件用于将其所在文件夹转化为 Python 包。然而在 Python 3 中，通过使用","0:\"%23333333\""],[20,"隐式命名空间包","0:\"%23333333\"|16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F\""],[20,"（Implicit Namespace Package）"],[20,"，这些文件就不再需要了。","0:\"%23333333\""],[20,"\n\nsound/ Top-level package"],[20,"\n","36:0"],[20," __init__.py Initialize the sound package"],[20,"\n","36:0"],[20," formats/ Subpackage for file format conversions"],[20,"\n","36:0"],[20," wavread.py"],[20,"\n","36:0"],[20," wavwrite.py"],[20,"\n","36:0"],[20," aiffread.py"],[20,"\n","36:0"],[20," aiffwrite.py"],[20,"\n","36:0"],[20," auread.py"],[20,"\n","36:0"],[20," auwrite.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," effects/ Subpackage for sound effects"],[20,"\n","36:0"],[20," echo.py"],[20,"\n","36:0"],[20," surround.py"],[20,"\n","36:0"],[20," reverse.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20," filters/ Subpackage for filters"],[20,"\n","36:0"],[20," equalizer.py"],[20,"\n","36:0"],[20," vocoder.py"],[20,"\n","36:0"],[20," karaoke.py"],[20,"\n","36:0"],[20," ..."],[20,"\n","36:0"],[20,"\n注：官方文档"],[20,"PEP 420 Specification","16:\"https%3A%2F%2Fwww.python.org%2Fdev%2Fpeps%2Fpep-0420%2F%23specification\""],[20," 指出，对于一些常规的包，__init__.py 仍然是需要的，如果将该文件删除，就会把该 Python 包变成一个本地的命名空间包，这将会产生一些额外的限制，具体可以参考这篇"],[20,"文档","16:\"https%3A%2F%2Fpackaging.python.org%2Fguides%2Fpackaging-namespace-packages%2F%23native-namespace-packages\""],[20,"。\n总结"],[20,"\n","32:1"],[20,"Python3 中的新功能远远不止如此，大家可以进一步进行探索。本文列出的内容只是一些实用的功能，希望能够对大家有所帮助。\n\n文中的代码链接："],[20,"\n","blockquote:true"],[20,"https://github.com/Weenkus/DataWhatNow-Codes/blob/master/things_you_are_probably_not_using_in_python_3_but_should/python%203%20examples.ipynb"],[20,"\n","blockquote:true"]]" style="font-size: 15px; color: rgb(63, 63, 63); letter-spacing: 1px;">

user = "Jane Doe"

复制代码

action = "buy"

复制代码

[/code][code]log_message = 'User {} has logged in and did an action {}.'.format(

复制代码

user,

复制代码

action

复制代码

)

复制代码

[/code][code]print(log_message)

复制代码

# User Jane Doe has logged in and did an action bu

复制代码

除了 format 之外， Python 3 还提供了一个更加灵活的方法来处理字符串，那就是 f-string 。如下所示，我们用 f-string 来和实现上面代码相同的功能：

user = "Jane Doe"

复制代码

action = "buy"

复制代码

[/code][code]log_message = f'User {user} has logged in and did an action {action}.'

复制代码

print(log_message)

复制代码

# User Jane Doe has logged in and did an action buy.

复制代码

Pathlib (3.4+)
如果需要处理文件路径，我们可以使用 Python 3 中的 pathlib 库，使对文件路径的操作更加便捷。如果你对 pathlib 存在疑惑，可以参考这篇文章。下面提供了一个代码示例：

from pathlib import Path

复制代码

[/code][code]root = Path('post_sub_folder')

复制代码

print(root)

复制代码

# post_sub_folder

复制代码

[/code][code]path = root / 'happy_user'

复制代码

# Make the path absolute

复制代码

print(path.resolve())

复制代码

# /home/weenkus/Workspace/Projects/DataWhatNow-Codes/how_your_python3_should_look_like/post_sub_folder/happy_user

复制代码

类型提示 (3.5+)

静态类型与动态类型是软件工程中的一个热门话题，Python 3 提供了支持 type hinting（类型提示）的方法，下面提供了一个示例：

def sentence_has_animal(sentence: str) -> bool:

复制代码

return "animal" in sentence

复制代码

sentence_has_animal("Donald had a farm without animals")

复制代码

[/code][code]# True

复制代码

枚举 (3.4+)

Python 3 中的 Enum 类支持枚举功能，可以使我们的程序变得更加简洁。 Enum 是一种便捷的变量列表的打包方式，使用该方法能够避免多个变量在代码各处分布，使代码显得杂乱无章。

from enum import Enum, auto

复制代码

[/code][code]class Monster(Enum):

复制代码

ZOMBIE = auto()

复制代码

WARRIOR = auto()

复制代码

BEAR = auto()

复制代码

print(Monster.ZOMBIE)

复制代码

# Monster.ZOMBIE

复制代码

枚举是一个符号集合，每个符号都和唯一的变量对应。通过使用枚举，我们可以通过符号标识来比较各个成员，我们还可以对枚举本身进行迭代。

https://docs.python.org/3/library/enum.html

for monster in Monster:

复制代码

print(monster)

复制代码

# Monster.ZOMBIE

复制代码

# Monster.WARRIOR

复制代码

# Monster.BEAR

复制代码

内置的 LRU 缓存（3.2+）

在现如今使用的硬件和软件中，缓存无所不在。Python 3 通过 lru_cache 来使用 LRU (Least Recently Used) 缓存。

下面的代码定义了一个斐波拉契函数，由于该函数的运算需要多次递归，每次递归都会执行相同的工作，因此使用缓存能够加速它的计算。

import time

复制代码

[/code][code]def fib(number: int) -> int:

复制代码

if number == 0: return 0

复制代码

if number == 1: return 1

复制代码

return fib(number-1) + fib(number-2)

复制代码

start = time.time()

复制代码

fib(40)

复制代码

print(f'Duration: {time.time() - start}s')

复制代码

# Duration: 30.684099674224854s

复制代码

我们可以使用 lru_cache 来优化该运算。这种优化极技术称为 memoization ，它能够把执行时间从几秒缩减到几纳秒。

from functools import lru_cache

复制代码

[/code][code]@lru_cache(maxsize=512)

复制代码

def fib_memoization(number: int) -> int:

复制代码

if number == 0: return 0

复制代码

if number == 1: return 1

复制代码

return fib_memoization(number-1) + fib_memoization(number-2)

复制代码

start = time.time()

复制代码

fib_memoization(40)

复制代码

print(f'Duration: {time.time() - start}s')

复制代码

[/code][code]# Duration: 6.866455078125e-05s

复制代码

扩展的可迭代解析功能（3.0+）

这里不做详细解释，直接贴出示例代码，具体解决可以参考这篇文档。

head, *body, tail = range(5)

复制代码

print(head, body, tail)

复制代码

# 0 [1, 2, 3] 4

复制代码

[/code][code]

复制代码

py, filename, *cmds = "python3.7 script.py -n 5 -l 15".split()

复制代码

print(py)

复制代码

print(filename)

复制代码

print(cmds)

复制代码

# python3.7

复制代码

# script.py

复制代码

# ['-n', '5', '-l', '15']

复制代码

[/code][code]

复制代码

first, _, third, *_ = range(10)

复制代码

print(first, third)

复制代码

# 0

复制代码

数据类（3.7+）
Python 3 引入了数据类（data class）。其装饰器会自动生成特征方法，例如 __init__() 和 __repr()__，这能够帮助减少样本代码的数量。在官方文档中，它们被称作 “具有默认值的可变命名元组” 。

class Armor:

复制代码

def __init__(self, armor: float, description: str, level: int = 1):

复制代码

self.armor = armor

复制代码

self.level = level

复制代码

self.description = description

复制代码

def power(self) -> float:

复制代码

return self.armor * self.level

复制代码

armor = Armor(5.2, "Common armor.", 2)

复制代码

armor.power()

复制代码

# 10.4

复制代码

[/code][code]print(armor)

复制代码

#

复制代码

接来下我们使用数据类来实现上面的代码：

from dataclasses import dataclass

复制代码

[/code][code]@dataclass

复制代码

class Armor:

复制代码

armor: float

复制代码

description: str

复制代码

level: int = 1

复制代码

def power(self) -> float:

复制代码

return self.armor * self.level

复制代码

armor = Armor(5.2, "Common armor.", 2)

复制代码

armor.power()

复制代码

# 10.4

复制代码

[/code][code]print(armor)

复制代码

# Armor(armor=5.2, description='Common armor.', level=2

复制代码

隐式命名空间包（3.3+）

构建Python代码有很多方法，其中一种就是在包（packages）中进行构建（即包含一个 __init__.py 文件的文件夹）。以下示例由 Python 的官方文档提供：

sound/ Top-level package

复制代码

__init__.py Initialize the sound package

复制代码

formats/ Subpackage for file format conversions

复制代码

__init__.py

复制代码

wavread.py

复制代码

wavwrite.py

复制代码

aiffread.py

复制代码

aiffwrite.py

复制代码

auread.py

复制代码

auwrite.py

复制代码

...

复制代码

effects/ Subpackage for sound effects

复制代码

__init__.py

复制代码

echo.py

复制代码

surround.py

复制代码

reverse.py

复制代码

...

复制代码

filters/ Subpackage for filters

复制代码

__init__.py

复制代码

equalizer.py

复制代码

vocoder.py

复制代码

karaoke.py

复制代码

...

复制代码

在 Python 2 中，上面的每个文件夹都必须有一个 __init__.py 文件，该文件用于将其所在文件夹转化为 Python 包。然而在 Python 3 中，通过使用隐式命名空间包（Implicit Namespace Package:https://www.python.org/dev/peps/pep-0420/），这些文件就不再需要了。

sound/ Top-level package

复制代码

__init__.py Initialize the sound package

复制代码

formats/ Subpackage for file format conversions

复制代码

wavread.py

复制代码

wavwrite.py

复制代码

aiffread.py

复制代码

aiffwrite.py

复制代码

auread.py

复制代码

auwrite.py

复制代码

...

复制代码

effects/ Subpackage for sound effects

复制代码

echo.py

复制代码

surround.py

复制代码

reverse.py

复制代码

...

复制代码

filters/ Subpackage for filters

复制代码

equalizer.py

复制代码

vocoder.py

复制代码

karaoke.py

复制代码

...

复制代码

注：官方文档PEP 420 Specification 指出，对于一些常规的包，__init__.py 仍然是需要的，如果将该文件删除，就会把该 Python 包变成一个本地的命名空间包，这将会产生一些额外的限制，具体可以参考这篇文档。

总结

Python3 中的新功能远远不止如此，大家可以进一步进行探索。本文列出的内容只是一些实用的功能，希望能够对大家有所帮助。

文中的代码链接：
https://github.com/Weenkus/DataWhatNow-Codes/blob/master/things_you_are_probably_not_using_in_python_3_but_should/python%203%20examples.ipynb

原文链接：
https://datawhatnow.com/things-you-are-probably-not-using-in-python-3-but-should/

（*本文由Python大本营编译，转载请联系微信1092722531）

◆
CTA核心技术及应用峰会
◆
5月25-27日，由中国IT社区CSDN与数字经济人才发展中心联合主办的第一届CTA核心技术及应用峰会将在杭州国际博览中心隆重召开，峰会将围绕人工智能领域，邀请技术领航者，与开发者共同探讨机器学习和知识图谱的前沿研究及应用。

更多重磅嘉宾请识别海报二维码查看。目前会议8折预售票抢购中，点击阅读原文即刻抢购。添加小助手微信15101014297，备注“CTA”，了解票务以及会务详情。

点击阅读原文，了解「CTA核心技术及应用峰会」