赞
踩
在 Python 处理文件的时候我们使用 with 关键词来进行文件的资源关闭,但是并不是只有文件操作才能使用 with 语句。今天就让我们一起学习 Python 中的上下文管理 contextlib。
上下文,简而言之,就是程式所执行的环境状态,或者说程式运行的情景。既然提及上下文,就不可避免的涉及 Python 中关于上下文的魔法。上下文管理器(context manager)是 Python2.5 开始支持的一种语法,用于规定某个对象的使用范围。一旦进入或者离开该使用范围,会有特殊操作被调用。它的语法形式是 with…as…,主要应用场景资源的创建和释放。例如,文件就支持上下文管理器,可以确保完成文件读写后关闭文件句柄。
with open('password.txt', 'wt') as f: f.write('contents go here')# 文件会自动关闭'password.txt', 'wt') as f: f.write('contents go here')# 文件会自动关闭
with 方法的实现涉及到两个魔法函数__enter__和__exit__。
执行流进入 with 中的代码块时会执行__enter__方法,它会返回在这个上下文中使用的一个对象。执行流离开 with 块时,则调用这个上下文管理器的__exit__方法来清理所使用的资源。
class Context: def __init__(self): print("int __init__") def __enter__(self): print("int __enter__") def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__")if __name__ == '__main__': with Context(): print('start with ') def __init__(self): print("int __init__") def __enter__(self): print("int __enter__") def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__")if __name__ == '__main__': with Context(): print('start with ')
输出
int __init__int __enter__start withtin __exit__int __enter__start withtin __exit__
相对于使用 try:finally 块,使用 with 语句代码看起来更紧凑,with 代码块执行的时候总会调用__exit__方法,即使出现了异常。
如果给 with 语句的 as 子句指定一个别名,那么__enter__方法可以返回与这个名关联的任何对象。
import requestsclass Request: def __init__(self): self.session = requests.session() def get(self, url, headers=None): if headers is None: headers = {} response = self.session.get(url) return responseclass Context: def __init__(self): print("int __init__") def __enter__(self): print("int __enter__") return Request() def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__")if __name__ == '__main__': with Context() as t: req = t.get("https://wwww.baidu.com") print(req.text)class Request: def __init__(self): self.session = requests.session() def get(self, url, headers=None): if headers is None: headers = {} response = self.session.get(url) return responseclass Context: def __init__(self): print("int __init__") def __enter__(self): print("int __enter__") return Request() def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__")if __name__ == '__main__': with Context() as t: req = t.get("https://wwww.baidu.com") print(req.text)
这里的 t 就是__enter__方法返回的 Request 对象的实例,然后我们可以调用该实例的一些方法。
如果上下文中出现异常,可以通过修改__exit__方法,如果返回值为 true,则可以把异常打印出来,如果为 false 则会抛出异常。
class Context: def __init__(self,flag): self.flag = flag def __enter__(self): print("int __enter__") return self def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__") print(f"{exc_type=}") #Python3.8 的语法等价于 f"exc_type={exc_type}" print(f"{exc_val=}") print(f"{exc_val=}") return self.flagif __name__ == '__main__': with Context(True) as t: raise RuntimeError() print("------华丽的分割线---------") with Context(False) as t: raise RuntimeError() def __init__(self,flag): self.flag = flag def __enter__(self): print("int __enter__") return self def __exit__(self, exc_type, exc_val, exc_tb): print("in __exit__") print(f"{exc_type=}") #Python3.8 的语法等价于 f"exc_type={exc_type}" print(f"{exc_val=}") print(f"{exc_val=}") return self.flagif __name__ == '__main__': with Context(True) as t: raise RuntimeError() print("------华丽的分割线---------") with Context(False) as t: raise RuntimeError()
输出
int __enter__in __exit__exc_type=<class 'RuntimeError'>exc_val=RuntimeError()exc_val=RuntimeError()------华丽的分割线---------int __enter__in __exit__exc_type=<class 'RuntimeError'>exc_val=RuntimeError()exc_val=RuntimeError()Traceback (most recent call last): File "demo.py", line 26, in <module> raise RuntimeError()RuntimeErrorexc_type=<class 'RuntimeError'>exc_val=RuntimeError()exc_val=RuntimeError()------华丽的分割线---------int __enter__in __exit__exc_type=<class 'RuntimeError'>exc_val=RuntimeError()exc_val=RuntimeError()Traceback (most recent call last): File "demo.py", line 26, in <module> raise RuntimeError()RuntimeError
可以看到__exit__方法接收一些参数,其中包含 with 块中产生的异常的详细信息。
通过继承 contextlib 里面的 ContextDecorator 类,实现对常规上下文管理器类的支持,其不仅可以作为上下文管理器,也可以作为函数修饰符。
import contextlibclass Context(contextlib.ContextDecorator): def __init__(self, how_used): self.how_used = how_used print(f'__init__({how_used})') def __enter__(self): print(f'__enter__({self.how_used})') return self def __exit__(self, exc_type, exc_val, exc_tb): print(f'__exit__({self.how_used})')@Context('这是装饰器方式')def func(message): print(message)print("---------------")func('作为装饰器运行')print("\n-------华丽的分割线--------\n")with Context('上下文管理器方式'): print('emmmm')class Context(contextlib.ContextDecorator): def __init__(self, how_used): self.how_used = how_used print(f'__init__({how_used})') def __enter__(self): print(f'__enter__({self.how_used})') return self def __exit__(self, exc_type, exc_val, exc_tb): print(f'__exit__({self.how_used})')@Context('这是装饰器方式')def func(message): print(message)print("---------------")func('作为装饰器运行')print("\n-------华丽的分割线--------\n")with Context('上下文管理器方式'): print('emmmm')
看一下 ContextDecorator 的源码就了解了
class ContextDecorator(object): "A base class or mixin that enables context managers to work as decorators." def _recreate_cm(self): """Return a recreated instance of self. Allows an otherwise one-shot context manager like _GeneratorContextManager to support use as a decorator via implicit recreation. This is a private interface just for _GeneratorContextManager. See issue #11647 for details. """ return self def __call__(self, func): #将类变为可调用对象 @wraps(func) def inner(*args, **kwds): with self._recreate_cm(): return func(*args, **kwds) return inner "A base class or mixin that enables context managers to work as decorators." def _recreate_cm(self): """Return a recreated instance of self. Allows an otherwise one-shot context manager like _GeneratorContextManager to support use as a decorator via implicit recreation. This is a private interface just for _GeneratorContextManager. See issue #11647 for details. """ return self def __call__(self, func): #将类变为可调用对象 @wraps(func) def inner(*args, **kwds): with self._recreate_cm(): return func(*args, **kwds) return inner
有时候我们的代码只有很少的上下文要管理,此时再使用上面的形式写出 with 相关的魔法函数
就显得比较啰嗦了,在这种情况下,我们可以使用 contextmanager 修饰符将一个生成器转换为上下文管理器。
import contextlib@contextlib.contextmanagerdef make_context(): print("enter make_context") try: yield {} except RuntimeError as err: print(f"{err=}")print("Normal")with make_context() as value: print("in with")print("RuntimeError")with make_context() as value: raise RuntimeError("runtimeerror")print("Else Error")with make_context() as value: raise ZeroDivisionError("0 不能作为分母")@contextlib.contextmanagerdef make_context(): print("enter make_context") try: yield {} except RuntimeError as err: print(f"{err=}")print("Normal")with make_context() as value: print("in with")print("RuntimeError")with make_context() as value: raise RuntimeError("runtimeerror")print("Else Error")with make_context() as value: raise ZeroDivisionError("0 不能作为分母")
输出结果
enter make_contextin withexistingRuntimeError:enter make_contexterr=RuntimeError('runtimeerror')existingElse Error:enter make_contextexistingTraceback (most recent call last): File "demo.py", line 24, in <module> raise ZeroDivisionError("0 不能作为除数")ZeroDivisionError: 0 不能作为除数in withexistingRuntimeError:enter make_contexterr=RuntimeError('runtimeerror')existingElse Error:enter make_contextexistingTraceback (most recent call last): File "demo.py", line 24, in <module> raise ZeroDivisionError("0 不能作为除数")ZeroDivisionError: 0 不能作为除数
需要主要的是这个函数必须是个装饰器
被装饰器装饰的函数分为三部分:
with 语句中的代码块执行前执行函数中 yield 之前代码
yield 返回的内容复制给 as 之后的变量
with 代码块执行完毕后执行函数中 yield 之后的代码
再简单理解就是
yield 前半段用来表示__enter__()
yield 后半段用来表示__exit__()
在这里,我们从 contextlib 模块中引入 contextmanager,然后装饰我们所定义的 make_context 函数。这就允许我们使用 Python 的 with 语句来调用 make_context 函数。在函数中通过 yield 一个空字典,将其传递出去,最终主调函数可以使用它。
一旦 with 语句结束,控制就会返回给 make_context 函数,它继续执行 yield 语句后面的代码,这个最终会执行 finally 语句打印 existing。如果我们遇到了 RuntimeError 错误,它就会被捕获,最终 finally 语句依然会打印 existing。
因为 contextmanager 继承自 ContextDecorator,所以也可以被用作函数修饰符
import contextlib@contextlib.contextmanagerdef make_context(): print("enter make_context") try: yield {} except RuntimeError as err: print(f"{err=}") finally: print("existing")@make_context()def normal(): print("in normal")@make_context()def raise_error(err): raise errif __name__ == '__main__': print("Normal:") normal() print("RuntimeError:") raise_error(RuntimeError("runtime error")) print("Else Error") raise_error(ValueError("value error"))@contextlib.contextmanagerdef make_context(): print("enter make_context") try: yield {} except RuntimeError as err: print(f"{err=}") finally: print("existing")@make_context()def normal(): print("in normal")@make_context()def raise_error(err): raise errif __name__ == '__main__': print("Normal:") normal() print("RuntimeError:") raise_error(RuntimeError("runtime error")) print("Else Error") raise_error(ValueError("value error"))
输出
Normal:enter make_contextin normalexistingRuntimeError:enter make_contexterr=RuntimeError('runtime error')existingElse Errorenter make_contextexistingTraceback (most recent call last): File "demo.py", line 33, in <module> raise_error(ValueError("value error")) File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/contextlib.py", line 75, in inner return func(*args, **kwds) File "demo.py", line 21, in raise_error raise errValueError: value errorin normalexistingRuntimeError:enter make_contexterr=RuntimeError('runtime error')existingElse Errorenter make_contextexistingTraceback (most recent call last): File "demo.py", line 33, in <module> raise_error(ValueError("value error")) File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/contextlib.py", line 75, in inner return func(*args, **kwds) File "demo.py", line 21, in raise_error raise errValueError: value error
这是 contextmanager 原理:
1、因为 func()已经是个生成器了嘛,所以运行enter_()的时候,contextmanager 调用 self.gen.next()会跑到 func 的 yield 处,停住挂起,这个时候已经有了 t1=time.time() 2、然后运行 with 语句体里面的语句,也就是 a+b=300 3、跑完后运行exit()的时候,contextmanager 调用 self.gen.next()会从 func 的 yield 的下一句开始一直到结束。这个时候有了 t2=time.time(),t2-t1 从而实现了统计 cost_time 的效果,完美。
源码
class GeneratorContextManager(object): """Helper for @contextmanager decorator.""" def __init__(self, gen): self.gen = gen def __enter__(self): try: return self.gen.next() except StopIteration: raise RuntimeError("generator didn't yield") def __exit__(self, type, value, traceback): if type is None: try: self.gen.next() except StopIteration: return else: raise RuntimeError("generator didn't stop") else: if value is None: # Need to force instantiation so we can reliably # tell if we get the same exception back value = type() try: self.gen.throw(type, value, traceback) raise RuntimeError("generator didn't stop after throw()") except StopIteration, exc: return exc is not value except: if sys.exc_info()[1] is not value: raisedef contextmanager(func): @wraps(func) def helper(*args, **kwds): return GeneratorContextManager(func(*args, **kwds)) return helper """Helper for @contextmanager decorator.""" def __init__(self, gen): self.gen = gen def __enter__(self): try: return self.gen.next() except StopIteration: raise RuntimeError("generator didn't yield") def __exit__(self, type, value, traceback): if type is None: try: self.gen.next() except StopIteration: return else: raise RuntimeError("generator didn't stop") else: if value is None: # Need to force instantiation so we can reliably # tell if we get the same exception back value = type() try: self.gen.throw(type, value, traceback) raise RuntimeError("generator didn't stop after throw()") except StopIteration, exc: return exc is not value except: if sys.exc_info()[1] is not value: raisedef contextmanager(func): @wraps(func) def helper(*args, **kwds): return GeneratorContextManager(func(*args, **kwds)) return helper
并不是所有的类都支持上下文管理器的 API,有一些遗留的类会使用一个 close 方法。
为了确保关闭句柄,需要使用 closing 为他创建一个上文管理器。
该模块主要是解决不能支持上下文管理器的类。
import contextlibclass Http(): def __init__(self): print("int init:") self.session = "open" def close(self): """ 关闭的方法必须叫 close """ print("in close:") self.session = "close"if __name__ == '__main__': with contextlib.closing(Http()) as http: print(f"inside session value:{http.session}") print(f"outside session value:{http.session}") with contextlib.closing(Http()) as http: print(f"inside session value:{http.session}") raise EnvironmentError("EnvironmentError") print(f"outside session value:{http.session}")class Http(): def __init__(self): print("int init:") self.session = "open" def close(self): """ 关闭的方法必须叫 close """ print("in close:") self.session = "close"if __name__ == '__main__': with contextlib.closing(Http()) as http: print(f"inside session value:{http.session}") print(f"outside session value:{http.session}") with contextlib.closing(Http()) as http: print(f"inside session value:{http.session}") raise EnvironmentError("EnvironmentError") print(f"outside session value:{http.session}")
输出
int init:inside session value:openin close:outside session value:closeint init:inside session value:openin close:Traceback (most recent call last): File "demo.py", line 23, in <module> raise EnvironmentError("EnvironmentError")OSError: EnvironmentErrorinside session value:openin close:outside session value:closeint init:inside session value:openin close:Traceback (most recent call last): File "demo.py", line 23, in <module> raise EnvironmentError("EnvironmentError")OSError: EnvironmentError
可以看到即使程序出现了错误,最后也会执行 close 方法的内容。
看一眼下面的源码就知道 closing 干嘛了
class closing(object): """Context to automatically close something at the end of a block. Code like this: with closing(<module>.open(<arguments>)) as f: <block> is equivalent to this: f = <module>.open(<arguments>) try: <block> finally: f.close() """ def __init__(self, thing): self.thing = thing def __enter__(self): return self.thing def __exit__(self, *exc_info): self.thing.close() """Context to automatically close something at the end of a block. Code like this: with closing(<module>.open(<arguments>)) as f: <block> is equivalent to this: f = <module>.open(<arguments>) try: <block> finally: f.close() """ def __init__(self, thing): self.thing = thing def __enter__(self): return self.thing def __exit__(self, *exc_info): self.thing.close()
这个 contextlib.closing()会帮它加上__enter__()和__exit__(),使其满足 with 的条件。然后 exit 里执行的就是对应类的 close 方法。
contextlib.suppress(*exceptions)
另一个工具就是在 Python 3.4 中加入的 suppress 类。这个上下文管理工具背后的理念就是它可以禁止任意数目的异常。假如我们想忽略 FileNotFoundError 异常。如果你书写了如下的上下文管理器,那么它不会正常运行。
>>> with open("1.txt") as fobj:... for line in fobj:... print(line)...Traceback (most recent call last): File "<stdin>", line 1, in <module>FileNotFoundError: [Errno 2] No such file or directory: '1.txt'"1.txt") as fobj:... for line in fobj:... print(line)...Traceback (most recent call last): File "<stdin>", line 1, in <module>FileNotFoundError: [Errno 2] No such file or directory: '1.txt'
正如你所看到的,这个上下文管理器没有处理这个异常,如果你想忽略这个错误,你可以按照如下方式来做,
>>> from contextlib import suppress>>> with suppress(FileNotFoundError):... with open("1.txt") as fobj:... for line in fobj:... print(line)import suppress>>> with suppress(FileNotFoundError):... with open("1.txt") as fobj:... for line in fobj:... print(line)
在这段代码中,我们引入 suppress,然后将我们要忽略的异常传递给它,在这个例子中,就是 FileNotFoundError。如果你想运行这段代码,你将会注意到,文件不存在时,什么事情都没有发生,也没有错误被抛出。请注意,这个上下文管理器是可重用的。
在编程中如果频繁的修改数据库, 一味的使用类似 try:… except..: rollback() raise e 其实是不太好的.
try: gift = Gift() gift.isbn = isbn ... db.session.add(gift) db.session.commit() except Exception as e: db.session.rollback() raise e gift = Gift() gift.isbn = isbn ... db.session.add(gift) db.session.commit() except Exception as e: db.session.rollback() raise e
为了达到使用 with 语句的目的, 我们可以重写 db 所属的类:
from flask_sqlalchemy import SQLAlchemy as _SQLALchemyclass SQLAlchemy(_SQLALchemy): @contextmanager def auto_commit(self): try: yield self.session.commit() except Exception as e: db.session.rollback() raise eimport SQLAlchemy as _SQLALchemyclass SQLAlchemy(_SQLALchemy): @contextmanager def auto_commit(self): try: yield self.session.commit() except Exception as e: db.session.rollback() raise e
这时候, 在执行数据的修改的时候便可以:
with db.auto_commit(): gift = Gift() gift.isbn = isbndb.session.add(gift) db.session.add(gift)with db.auto_commit(): user = User() user.set_attrs(form.data) db.session.add(user) gift.isbn = isbndb.session.add(gift) db.session.add(gift)with db.auto_commit(): user = User() user.set_attrs(form.data) db.session.add(user)
上下文管理器很有趣,也很方便。我经常在自动测试中使用它们,例如,打开和关闭对话。现在,你应该可以使用 Python 内置的工具去创建你的上下文管理器。你还可以继续阅读 Python 关于 contextlib 的文档,那里有很多本文没有覆盖到的知识。
这篇文章其实主要为了后面的异步上文管理器使用的。
https://www.cnblogs.com/zhbzz2007/p/6158125.html
https://blog.csdn.net/emaste_r/article/details/78105713
https://blog.csdn.net/weixin_42359464/article/details/80742387#three
https://docs.python.org/zh-cn/3/library/contextlib.html?highlight=contextlib#module-contextlib
点击阅读原文获取Python 爬虫面试题 170 道:2019 版
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。