Diff for "Decorators"

Differences between revisions 1 and 4 (spanning 3 versions)

Apart from decorators in the standard library, there are also lots of PyPI packages with convenient, interesting or novel use cases:

Category	Decorator	Summary
builtin	@property	Turn function into propery accessor, also: `@property.setter`
builtin	@classsmethod	Use `cls` instead of `self`
builtin	@staticmethod	Defines methods without self handle
stdlib	@functools.wraps	Retain modicum of meta infos (`__doc__`) of wrapped function
stdlib	@functools.singledispatch	Allows for divergent implementations per @f.register()
stdlib/backports	@functools.lru_cache	Wrap computationally expensive functions
stdlib	@functools.cmp_to_key	Transform an old-style comparison function to a key function.
stdlib	@functools.total_ordering	Adds any of the missing lt,le,eq,ge,gt comparison dunder methods
stdlib	@functools.partial	Prepopulate a couple of function arguments
stdlib	@typing.overload	Singledispatch based on function parameter or return types
stdlib	@contextlib.contextmanager	Uses generator protocol `yield` to turn function into `with`-able context mananger
stdlib	@dataclasses.dataclass	Adds various “dunder” methods to the class to streamline property access
stdlib	@atexit.register	Run as shutdown function
stdlib	@abc.ABCMeta.register	Turn into abstract base class
stdlib	@enum.unique	Guarantee unique members in enum class
stdlib	@xmlrpc.register_function	...
stdlib	@sys.settrace	Register function as debugger callback
asyncio	@coroutine	Outdated scheme for `async def`, deprecated in 3.8
asyncio	@throttle(2,1)	Reduce invocations per timeframe
asyncio	@exceptionCatcher…	Log exceptions more instantly
asyncio	@timeit	Basic benchmarking
async	@async_process	make func async and execute in other process
threading	@threaded, @debug, @skip_if_env	Convenient threading wrapper, and testing wrappers
threading	@lock_or_wait	Aquire and release locks for functions
testing	@unittest.mock.patch	Override function w/ signature
testing	@wrapt.patch_function_wrapper	Override function w/ signature
debugging	@trycatch	Capture exceptions to the log
debugging	@log_calls	Record invocations with arguments
decorator	@decorator	Assists declaring signature-preserving and single dispatch decorators
decorator	@decorator_args	Ease parameterizing decorators themselvess
decorator	@named_decorator	Retain callees' function name to simplify tracebacks
decorator	@dec.decorator(before=cb,after=cb)	Craft trivial enter/exit decorator
decorator	@log_call, @intercept, @instead, @before	Simplifies decorator construnction/interactions, parameter handling
decorator	undecorated(fn)	reversible decoration
collect	@click.option("-d")	Argparse alternative
collect	@config_decorator.setting	Declarative dict structure for application settings
collect	@regex_decorator.listen	Register function callbacks for pattern matching
collect	@expose, @register	Plugin interface/registration
typing	@contract({"param":gt(3)})	Exhaustive function parameter constraints
typing	@pedantic, @trace, @dirty, …	Encourages cleaner and better documented code
typing	@Decorator	OO-style class/method parameters
typing	@mutable, @final, @auto_obj	decorators for str/repr, equality, immutability, and more
typing	@to_string, @data, @repeat	Reduces some common boilerplate for methods
typing	@DataAtrribute("desc")	.NET-like attributes
parameter	@curried	Convenient decorator-alternative to partial()
parameter	@attrs	object declarations
parameter	@arg.validate	simplifies parameter handling/assertion
flow handling	@__main__	Automatically run a function if invocation script
flow handling	@retry(times=2)	Reinvokes function a few times when encountering (temporary) exceptions, alt: pypi:retry-decorator
flow handling	@retry	Rerun function in case of exceptions
flow handling	@self	method chaining (fluent interfaces)
monkeypatching	@inject(modules…)	Shorthand assignment to modules/objects
monkeypatching	@monkeybiz.patch(fn)	Swap out function in object, retain reference to original

(Note: Not sure this is going anywhere. Relisting builtins is somewhat redundant, but on topic here. The focus is novel/interesting decorators in the wild. The categorization probably won't hold up; and probably going to split this up into sections.)

main

This decorator does not alter a function, but causes it to be executed if __name__ == '__main__'. This provides an experimental cleaner syntax to the traditional way of bootstrapping a python script. This should not be used on class methods.

The function gets a copied list of sys.argv arguments.

   1 def __main__(func):
   2 
   3     if __name__ == "__main__":
   4         import sys, os
   5         args = sys.argv[:]
   6         args[0] = os.path.abspath(args[0])
   7         func(args)

-  ⇤ ← Revision 1 as of 2005-03-02 17:58:05 → 
  Size: 578
  Editor: 160
  Comment:
+   ← Revision 4 as of 2021-04-06 09:53:39 → ⇥
  Size: 9623
  Editor: ProfMilki
  Comment: Expand with ~12 more decorator packages
-Deletions are marked like this.
+Additions are marked like this.
 Line 1:
-A page for useful (or potentially abusive?) decorator ideas.
+Apart from decorators in the standard library, there are also lots of PyPI packages with convenient, interesting or novel use cases:

|| '''Category'''   || '''Decorator'''      || '''Summary'''                                        ||
|| builtin          || [[https://docs.python.org/3/library/functions.html?highlight=property#property|@property]]  || Turn function into propery accessor, also: `@property.setter` ||
|| builtin          || [[https://docs.python.org/3/library/functions.html?highlight=property#classmethod|@classsmethod]]  || Use `cls` instead of `self` ||
|| builtin          || [[https://docs.python.org/3/library/functions.html?highlight=property#staticmethod|@staticmethod]] || Defines methods without self handle ||
|| stdlib           || [[https://docs.python.org/3/library/functools.html?highlight=functools#functools.wraps|@functools.wraps]]  || Retain modicum of meta infos (`__doc__`) of wrapped function   ||
|| stdlib           || [[https://docs.python.org/3/library/functools.html?highlight=functools#functools.singledispatch|@functools.singledispatch]]    || Allows for divergent implementations per @f.register()  ||
|| stdlib/backports || [[https://stackoverflow.com/questions/815110/is-there-a-decorator-to-simply-cache-function-return-values|@functools.lru_cache]]  || Wrap computationally expensive functions   ||
|| stdlib           || [[https://docs.python.org/3/library/functools.html?highlight=functools#functools.cmp_to_key|@functools.cmp_to_key]] || Transform an old-style comparison function to a key function.  ||
|| stdlib           || [[https://docs.python.org/3/library/functools.html?highlight=functools#functools.total_ordering|@functools.total_ordering]] || Adds any of the missing lt,le,eq,ge,gt comparison dunder methods  ||
|| stdlib           || [[https://docs.python.org/3/library/functools.html?highlight=functools#functools.partial|@functools.partial]] || Prepopulate a couple of function arguments   ||
|| stdlib           || [[https://python.readthedocs.io/en/stable/library/typing.html#typing.overload|@typing.overload]]  || Singledispatch based on function parameter or return types  ||
|| stdlib           || [[https://docs.python.org/3/library/contextlib.html?highlight=decorator#contextlib.contextmanager|@contextlib.contextmanager]] || Uses generator protocol `yield` to turn function into `with`-able context mananger  ||
|| stdlib           || [[https://docs.python.org/3/library/dataclasses.html?highlight=decorator#dataclasses.dataclass|@dataclasses.dataclass]] || Adds various “dunder” methods to the class to streamline property access  ||
|| stdlib           || [[https://docs.python.org/3/library/atexit.html?highlight=decorator#atexit.register|@atexit.register]]  || Run as shutdown function ||
|| stdlib           || [[https://docs.python.org/3/library/abc.html?highlight=decorator#abc.ABCMeta.register|@abc.ABCMeta.register]]  || Turn into abstract base class  ||
|| stdlib           || [[https://docs.python.org/3/library/enum.html?highlight=decorator#enum.unique|@enum.unique]]  || Guarantee unique members in enum class  ||
|| stdlib           || [[https://docs.python.org/3a/library/xmlrpc.server.html?highlight=decorator#xmlrpc.server.SimpleXMLRPCServer.register_function|@xmlrpc.register_function]]  || ...  ||
|| stdlib           || [[https://docs.python.org/3/library/sys.html#sys.settrace|@sys.settrace]]  || Register function as debugger callback ||
|| asyncio          || [[https://docs.python.org/3/library/asyncio-task.html?highlight=asyncio%20coroutine#asyncio.coroutine|@coroutine]]  || Outdated scheme for `async def`, deprecated in 3.8  ||
|| asyncio          || [[https://pypi.org/project/aiodecorator/|@throttle(2,1)]]    || Reduce invocations per timeframe    ||
|| asyncio          || [[https://hackernoon.com/python-async-decorator-to-reduce-debug-woes-nv2dg30q5|@exceptionCatcher…]]  || Log exceptions more instantly   ||
|| asyncio          || [[https://gist.github.com/Integralist/77d73b2380e4645b564c28c53fae71fb|@timeit]]  || Basic benchmarking   ||
|| async            || [[https://pypi.org/project/process-decorator/|@async_process]]  || make func async and execute in other process  ||
|| threading        || [[https://gitlab.com/chadgh/ornamentation|@threaded, @debug, @skip_if_env]]  || Convenient threading wrapper, and testing wrappers  ||
|| threading        || [[https://pypi.org/project/lockorator/|@lock_or_wait]]  || Aquire and release locks for functions  ||
|| testing          || [[https://docs.python.org/3/library/unittest.mock.html#unittest.mock.patch|@unittest.mock.patch]]  || Override function w/ signature  ||
|| testing          || [[http://blog.dscpl.com.au/2015/03/using-wrapt-to-support-testing-of.html|@wrapt.patch_function_wrapper]]  || Override function w/ signature  ||
|| debugging        || [[https://pypi.org/project/handy-decorators/|@trycatch ]]  || Capture exceptions to the log  ||
|| debugging        || [[https://pythonhosted.org/log_calls/intro.html|@log_calls]]  || Record invocations with arguments  ||
|| decorator        || [[https://pypi.org/project/decorator/|@decorator]]  || Assists declaring signature-preserving and single dispatch decorators  ||
|| decorator        || [[https://pypi.org/project/decorator-args/|@decorator_args]]  || Ease parameterizing decorators themselvess  ||
|| decorator        || [[https://github.com/Yelp/named_decorator|@named_decorator]]  || Retain callees' function name to simplify tracebacks  ||
|| decorator        || [[https://pypi.org/project/Easy_Decorator/|@dec.decorator(before=cb,after=cb)]]  || Craft trivial enter/exit decorator  ||
|| decorator        || [[https://pypi.org/project/pydecor/|@log_call, @intercept, @instead, @before]]  || Simplifies decorator construnction/interactions, parameter handling ||
|| decorator        || [[https://pypi.org/project/undecorated/|undecorated(fn)]]  || reversible decoration  ||
|| collect          || [[https://click.palletsprojects.com/en/7.x/commands/|@click.option("-d")]] || Argparse alternative ||
|| collect          || [[https://pypi.org/project/config-decorator/|@config_decorator.setting]] || Declarative dict structure for application settings ||
|| collect          || [[https://github.com/xtream1101/regex-decorator|@regex_decorator.listen]] || Register function callbacks for pattern matching ||
|| collect          || [[https://pypi.org/project/pyplugs/|@expose, @register]]  || Plugin interface/registration ||
|| typing           || [[https://pypi.org/project/contract-decorator/|@contract({"param":gt(3)})]] || Exhaustive function parameter constraints ||
|| typing           || [[https://pypi.org/project/pedantic/|@pedantic, @trace, @dirty, …]]  || Encourages cleaner and better documented code ||
|| typing           || [[https://pypi.org/project/classy-decorators/|@Decorator]]  || OO-style class/method parameters  ||
|| typing           || [[https://pypi.org/project/decorateme/|@mutable, @final, @auto_obj]]  ||  decorators for str/repr, equality, immutability, and more  ||
|| typing           || [[https://pypi.org/project/paprika/|@to_string, @data, @repeat]]  || Reduces some common boilerplate for methods  ||
|| typing           || [[https://pypi.org/project/pyAttributes/|@DataAtrribute("desc")]]  || .NET-like attributes ||
|| parameter        || [[https://pypi.org/project/simplecurry/|@curried]]  || Convenient decorator-alternative to partial()  ||
|| parameter        || [[https://www.attrs.org/en/stable/|@attrs]]  || object declarations ||
|| parameter        || [[https://pypi.org/project/python-args/|@arg.validate]]  || simplifies parameter handling/assertion ||
|| flow handling    || [[#A__main__|@__main__]]  || Automatically run a function if invocation script  ||
|| flow handling    || [[http://www.praddy.in/retry-decorator-whitelisted-exceptions/|@retry(times=2)]]  || Reinvokes function a few times when encountering (temporary) exceptions, alt: [[https://pypi.org/project/retry-decorator/|pypi:retry-decorator]]  ||
|| flow handling    || [[https://pypi.org/project/retryz/|@retry]]  || Rerun function in case of exceptions ||
|| flow handling    || [[https://pypi.org/project/self/|@self]]  || method chaining (fluent interfaces)  ||
|| monkeypatching   || [[https://fossil.include-once.org/modseccfg/wiki/@inject|@inject(modules…)]]  || Shorthand assignment to modules/objects   ||
|| monkeypatching   || [[https://github.com/theatlantic/python-monkey-business|@monkeybiz.patch(fn)]]  || Swap out function in object, retain reference to original  ||

(Note: Not sure this is going anywhere. Relisting builtins is somewhat redundant, but on topic here.
The focus is novel/interesting decorators in the wild. The categorization probably won't hold up;
and probably going to split this up into sections.)


=== Other decorator links ===

 * [[https://github.com/lord63/awesome-python-decorator|Awesome Python decorator collection (GitHub)]]
 * [[https://pypi.org/search/?q=decorator|PyPI decorator packages]] (gets interesting around page 10)

Page

User

Other decorator links

main