This page is meant to be a central repository of decorator code pieces, whether useful or not <wink>. It is NOT a page to discuss decorator syntax!
Feel free to add your suggestions. Please make sure example code conforms with PEP 8.
Contents
- Creating Well-Behaved Decorators / "Decorator decorator"
- Property Definition
- Memoize
- Alternate memoize as nested functions
- Alternate memoize as dict subclass
- Alternate memoize that stores cache between executions
- Cached Properties
- Retry
- Pseudo-currying
- Creating decorator with optional arguments
- Controllable DIY debug
- Easy adding methods to a class instance
- Counting function calls
- Alternate Counting function calls
- Generating Deprecation Warnings
- Smart deprecation warnings (with valid filenames, line numbers, etc.)
- Ignoring Deprecation Warnings
- Enable/Disable Decorators
- Easy Dump of Function Arguments
- Pre-/Post-Conditions
- Profiling/Coverage Analysis
- Line Tracing Individual Functions
- Synchronization
- Type Enforcement (accepts/returns)
- CGI method wrapper
- State Machine Implementaion
- C++/Java-keyword-like function decorators
- Different Decorator Forms
- Unimplemented function replacement
- Redirects stdout printing to python standard logging.
- Access control
- Events rising and handling
- Singleton
- Asynchronous Call
- Class method decorator using instance
- Another Retrying Decorator
- Logging decorator with specified logger (or default)
- Lazy Thunkify
- Aggregative decorators for generator functions
- Function Timeout
- Collect Data Difference Caused by Decorated Function
Creating Well-Behaved Decorators / "Decorator decorator"
Note: This is only one recipe. Others include inheritance from a standard decorator (link?), the functools @wraps decorator, and a factory function such as Michele Simionato's decorator module which even preserves signature information.
1 def simple_decorator(decorator):
2 '''This decorator can be used to turn simple functions
3 into well-behaved decorators, so long as the decorators
4 are fairly simple. If a decorator expects a function and
5 returns a function (no descriptors), and if it doesn't
6 modify function attributes or docstring, then it is
7 eligible to use this. Simply apply @simple_decorator to
8 your decorator and it will automatically preserve the
9 docstring and function attributes of functions to which
10 it is applied.'''
11 def new_decorator(f):
12 g = decorator(f)
13 g.__name__ = f.__name__
14 g.__doc__ = f.__doc__
15 g.__dict__.update(f.__dict__)
16 return g
17 # Now a few lines needed to make simple_decorator itself
18 # be a well-behaved decorator.
19 new_decorator.__name__ = decorator.__name__
20 new_decorator.__doc__ = decorator.__doc__
21 new_decorator.__dict__.update(decorator.__dict__)
22 return new_decorator
23
24 #
25 # Sample Use:
26 #
27 @simple_decorator
28 def my_simple_logging_decorator(func):
29 def you_will_never_see_this_name(*args, **kwargs):
30 print 'calling {}'.format(func.__name__)
31 return func(*args, **kwargs)
32 return you_will_never_see_this_name
33
34 @my_simple_logging_decorator
35 def double(x):
36 'Doubles a number.'
37 return 2 * x
38
39 assert double.__name__ == 'double'
40 assert double.__doc__ == 'Doubles a number.'
41 print double(155)
Property Definition
These decorators provide a readable way to define properties:
1 import sys
2
3 def propget(func):
4 locals = sys._getframe(1).f_locals
5 name = func.__name__
6 prop = locals.get(name)
7 if not isinstance(prop, property):
8 prop = property(func, doc=func.__doc__)
9 else:
10 doc = prop.__doc__ or func.__doc__
11 prop = property(func, prop.fset, prop.fdel, doc)
12 return prop
13
14 def propset(func):
15 locals = sys._getframe(1).f_locals
16 name = func.__name__
17 prop = locals.get(name)
18 if not isinstance(prop, property):
19 prop = property(None, func, doc=func.__doc__)
20 else:
21 doc = prop.__doc__ or func.__doc__
22 prop = property(prop.fget, func, prop.fdel, doc)
23 return prop
24
25 def propdel(func):
26 locals = sys._getframe(1).f_locals
27 name = func.__name__
28 prop = locals.get(name)
29 if not isinstance(prop, property):
30 prop = property(None, None, func, doc=func.__doc__)
31 else:
32 prop = property(prop.fget, prop.fset, func, prop.__doc__)
33 return prop
34
35 # These can be used like this:
36
37 class Example(object):
38
39 @propget
40 def myattr(self):
41 return self._half * 2
42
43 @propset
44 def myattr(self, value):
45 self._half = value / 2
46
47 @propdel
48 def myattr(self):
49 del self._half
Here's a way that doesn't require any new decorators:
1 class Example(object):
2 @apply # doesn't exist in Python 3
3 def myattr():
4 doc = '''This is the doc string.'''
5
6 def fget(self):
7 return self._half * 2
8
9 def fset(self, value):
10 self._half = value / 2
11
12 def fdel(self):
13 del self._half
14
15 return property(**locals())
16 #myattr = myattr() # works in Python 2 and 3
Yet another property decorator:
1 try:
2 # Python 2
3 import __builtin__ as builtins
4 except ImportError:
5 # Python 3
6 import builtins
7
8 def property(function):
9 keys = 'fget', 'fset', 'fdel'
10 func_locals = {'doc':function.__doc__}
11 def probe_func(frame, event, arg):
12 if event == 'return':
13 locals = frame.f_locals
14 func_locals.update(dict((k, locals.get(k)) for k in keys))
15 sys.settrace(None)
16 return probe_func
17 sys.settrace(probe_func)
18 function()
19 return builtins.property(**func_locals)
20
21 #====== Example =======================================================
22
23 from math import radians, degrees, pi
24
25 class Angle(object):
26 def __init__(self, rad):
27 self._rad = rad
28
29 @property
30 def rad():
31 '''The angle in radians'''
32 def fget(self):
33 return self._rad
34 def fset(self, angle):
35 if isinstance(angle, Angle):
36 angle = angle.rad
37 self._rad = float(angle)
38
39 @property
40 def deg():
41 '''The angle in degrees'''
42 def fget(self):
43 return degrees(self._rad)
44 def fset(self, angle):
45 if isinstance(angle, Angle):
46 angle = angle.deg
47 self._rad = radians(angle)
Memoize
Here's a memoizing class.
1 import collections
2 import functools
3
4 class memoized(object):
5 '''Decorator. Caches a function's return value each time it is called.
6 If called later with the same arguments, the cached value is returned
7 (not reevaluated).
8 '''
9 def __init__(self, func):
10 self.func = func
11 self.cache = {}
12 def __call__(self, *args):
13 if not isinstance(args, collections.Hashable):
14 # uncacheable. a list, for instance.
15 # better to not cache than blow up.
16 return self.func(*args)
17 if args in self.cache:
18 return self.cache[args]
19 else:
20 value = self.func(*args)
21 self.cache[args] = value
22 return value
23 def __repr__(self):
24 '''Return the function's docstring.'''
25 return self.func.__doc__
26 def __get__(self, obj, objtype):
27 '''Support instance methods.'''
28 return functools.partial(self.__call__, obj)
29
30 @memoized
31 def fibonacci(n):
32 "Return the nth fibonacci number."
33 if n in (0, 1):
34 return n
35 return fibonacci(n-1) + fibonacci(n-2)
36
37 print fibonacci(12)
Alternate memoize as nested functions
Here's a memoizing function that works on functions, methods, or classes, and exposes the cache publicly.
Here's a modified version that also respects kwargs.
Alternate memoize as dict subclass
This is an idea that interests me, but it only seems to work on functions:
1 class memoize(dict):
2 def __init__(self, func):
3 self.func = func
4
5 def __call__(self, *args):
6 return self[args]
7
8 def __missing__(self, key):
9 result = self[key] = self.func(*key)
10 return result
11
12 #
13 # Sample use
14 #
15
16 >>> @memoize
17 ... def foo(a, b):
18 ... return a * b
19 >>> foo(2, 4)
20 8
21 >>> foo
22 {(2, 4): 8}
23 >>> foo('hi', 3)
24 'hihihi'
25 >>> foo
26 {(2, 4): 8, ('hi', 3): 'hihihi'}
Alternate memoize that stores cache between executions
Additional information and documentation for this decorator is available on Github.
1 import pickle
2 import collections
3 import functools
4 import inspect
5 import os.path
6 import re
7 import unicodedata
8
9 class Memorize(object):
10 '''
11 A function decorated with @Memorize caches its return
12 value every time it is called. If the function is called
13 later with the same arguments, the cached value is
14 returned (the function is not reevaluated). The cache is
15 stored as a .cache file in the current directory for reuse
16 in future executions. If the Python file containing the
17 decorated function has been updated since the last run,
18 the current cache is deleted and a new cache is created
19 (in case the behavior of the function has changed).
20 '''
21 def __init__(self, func):
22 self.func = func
23 self.set_parent_file() # Sets self.parent_filepath and self.parent_filename
24 self.__name__ = self.func.__name__
25 self.set_cache_filename()
26 if self.cache_exists():
27 self.read_cache() # Sets self.timestamp and self.cache
28 if not self.is_safe_cache():
29 self.cache = {}
30 else:
31 self.cache = {}
32
33 def __call__(self, *args):
34 if not isinstance(args, collections.Hashable):
35 return self.func(*args)
36 if args in self.cache:
37 return self.cache[args]
38 else:
39 value = self.func(*args)
40 self.cache[args] = value
41 self.save_cache()
42 return value
43
44 def set_parent_file(self):
45 """
46 Sets self.parent_file to the absolute path of the
47 file containing the memoized function.
48 """
49 rel_parent_file = inspect.stack()[-1].filename
50 self.parent_filepath = os.path.abspath(rel_parent_file)
51 self.parent_filename = _filename_from_path(rel_parent_file)
52
53 def set_cache_filename(self):
54 """
55 Sets self.cache_filename to an os-compliant
56 version of "file_function.cache"
57 """
58 filename = _slugify(self.parent_filename.replace('.py', ''))
59 funcname = _slugify(self.__name__)
60 self.cache_filename = filename+'_'+funcname+'.cache'
61
62 def get_last_update(self):
63 """
64 Returns the time that the parent file was last
65 updated.
66 """
67 last_update = os.path.getmtime(self.parent_filepath)
68 return last_update
69
70 def is_safe_cache(self):
71 """
72 Returns True if the file containing the memoized
73 function has not been updated since the cache was
74 last saved.
75 """
76 if self.get_last_update() > self.timestamp:
77 return False
78 return True
79
80 def read_cache(self):
81 """
82 Read a pickled dictionary into self.timestamp and
83 self.cache. See self.save_cache.
84 """
85 with open(self.cache_filename, 'rb') as f:
86 data = pickle.loads(f.read())
87 self.timestamp = data['timestamp']
88 self.cache = data['cache']
89
90 def save_cache(self):
91 """
92 Pickle the file's timestamp and the function's cache
93 in a dictionary object.
94 """
95 with open(self.cache_filename, 'wb+') as f:
96 out = dict()
97 out['timestamp'] = self.get_last_update()
98 out['cache'] = self.cache
99 f.write(pickle.dumps(out))
100
101 def cache_exists(self):
102 '''
103 Returns True if a matching cache exists in the current directory.
104 '''
105 if os.path.isfile(self.cache_filename):
106 return True
107 return False
108
109 def __repr__(self):
110 """ Return the function's docstring. """
111 return self.func.__doc__
112
113 def __get__(self, obj, objtype):
114 """ Support instance methods. """
115 return functools.partial(self.__call__, obj)
116
117 def _slugify(value):
118 """
119 Normalizes string, converts to lowercase, removes
120 non-alpha characters, and converts spaces to
121 hyphens. From
122 http://stackoverflow.com/questions/295135/turn-a-string-into-a-valid-filename-in-python
123 """
124 value = unicodedata.normalize('NFKD', value).encode('ascii', 'ignore')
125 value = re.sub(r'[^\w\s-]', '', value.decode('utf-8', 'ignore'))
126 value = value.strip().lower()
127 value = re.sub(r'[-\s]+', '-', value)
128 return value
129
130 def _filename_from_path(filepath):
131 return filepath.split('/')[-1]
Cached Properties
1 #
2 # © 2011 Christopher Arndt, MIT License
3 #
4
5 import time
6
7 class cached_property(object):
8 '''Decorator for read-only properties evaluated only once within TTL period.
9
10 It can be used to create a cached property like this::
11
12 import random
13
14 # the class containing the property must be a new-style class
15 class MyClass(object):
16 # create property whose value is cached for ten minutes
17 @cached_property(ttl=600)
18 def randint(self):
19 # will only be evaluated every 10 min. at maximum.
20 return random.randint(0, 100)
21
22 The value is cached in the '_cache' attribute of the object instance that
23 has the property getter method wrapped by this decorator. The '_cache'
24 attribute value is a dictionary which has a key for every property of the
25 object which is wrapped by this decorator. Each entry in the cache is
26 created only when the property is accessed for the first time and is a
27 two-element tuple with the last computed property value and the last time
28 it was updated in seconds since the epoch.
29
30 The default time-to-live (TTL) is 300 seconds (5 minutes). Set the TTL to
31 zero for the cached value to never expire.
32
33 To expire a cached property value manually just do::
34
35 del instance._cache[<property name>]
36
37 '''
38 def __init__(self, ttl=300):
39 self.ttl = ttl
40
41 def __call__(self, fget, doc=None):
42 self.fget = fget
43 self.__doc__ = doc or fget.__doc__
44 self.__name__ = fget.__name__
45 self.__module__ = fget.__module__
46 return self
47
48 def __get__(self, inst, owner):
49 now = time.time()
50 try:
51 value, last_update = inst._cache[self.__name__]
52 if self.ttl > 0 and now - last_update > self.ttl:
53 raise AttributeError
54 except (KeyError, AttributeError):
55 value = self.fget(inst)
56 try:
57 cache = inst._cache
58 except AttributeError:
59 cache = inst._cache = {}
60 cache[self.__name__] = (value, now)
61 return value
Retry
Call a function which returns True/False to indicate success or failure. On failure, wait, and try the function again. On repeated failures, wait longer between each successive attempt. If the decorator runs out of attempts, then it gives up and returns False, but you could just as easily raise some exception.
1 import time
2 import math
3
4 # Retry decorator with exponential backoff
5 def retry(tries, delay=3, backoff=2):
6 '''Retries a function or method until it returns True.
7
8 delay sets the initial delay in seconds, and backoff sets the factor by which
9 the delay should lengthen after each failure. backoff must be greater than 1,
10 or else it isn't really a backoff. tries must be at least 0, and delay
11 greater than 0.'''
12
13 if backoff <= 1:
14 raise ValueError("backoff must be greater than 1")
15
16 tries = math.floor(tries)
17 if tries < 0:
18 raise ValueError("tries must be 0 or greater")
19
20 if delay <= 0:
21 raise ValueError("delay must be greater than 0")
22
23 def deco_retry(f):
24 def f_retry(*args, **kwargs):
25 mtries, mdelay = tries, delay # make mutable
26
27 rv = f(*args, **kwargs) # first attempt
28 while mtries > 0:
29 if rv is True: # Done on success
30 return True
31
32 mtries -= 1 # consume an attempt
33 time.sleep(mdelay) # wait...
34 mdelay *= backoff # make future wait longer
35
36 rv = f(*args, **kwargs) # Try again
37
38 return False # Ran out of tries :-(
39
40 return f_retry # true decorator -> decorated function
41 return deco_retry # @retry(arg[, ...]) -> true decorator
Pseudo-currying
(FYI you can use functools.partial() to emulate currying (which works even for keyword arguments))
1 class curried(object):
2 '''
3 Decorator that returns a function that keeps returning functions
4 until all arguments are supplied; then the original function is
5 evaluated.
6 '''
7
8 def __init__(self, func, *a):
9 self.func = func
10 self.args = a
11
12 def __call__(self, *a):
13 args = self.args + a
14 if len(args) < self.func.func_code.co_argcount:
15 return curried(self.func, *args)
16 else:
17 return self.func(*args)
18
19
20 @curried
21 def add(a, b):
22 return a + b
23
24 add1 = add(1)
25
26 print add1(2)
Creating decorator with optional arguments
1 import functools, inspect
2
3 def decorator(func):
4 ''' Allow to use decorator either with arguments or not. '''
5
6 def isFuncArg(*args, **kw):
7 return len(args) == 1 and len(kw) == 0 and (
8 inspect.isfunction(args[0]) or isinstance(args[0], type))
9
10 if isinstance(func, type):
11 def class_wrapper(*args, **kw):
12 if isFuncArg(*args, **kw):
13 return func()(*args, **kw) # create class before usage
14 return func(*args, **kw)
15 class_wrapper.__name__ = func.__name__
16 class_wrapper.__module__ = func.__module__
17 return class_wrapper
18
19 @functools.wraps(func)
20 def func_wrapper(*args, **kw):
21 if isFuncArg(*args, **kw):
22 return func(*args, **kw)
23
24 def functor(userFunc):
25 return func(userFunc, *args, **kw)
26
27 return functor
28
29 return func_wrapper
Example:
1 @decorator
2 def apply(func, *args, **kw):
3 return func(*args, **kw)
4
5 @decorator
6 class apply:
7 def __init__(self, *args, **kw):
8 self.args = args
9 self.kw = kw
10
11 def __call__(self, func):
12 return func(*self.args, **self.kw)
13
14 #
15 # Usage in both cases:
16 #
17 @apply
18 def test():
19 return 'test'
20
21 assert test == 'test'
22
23 @apply(2, 3)
24 def test(a, b):
25 return a + b
26
27 assert test is 5
Note: There is only one drawback: wrapper checks its arguments for single function or class. To avoid wrong behavior you can use keyword arguments instead of positional, e.g.:
Controllable DIY debug
(Other hooks could be similarly added. Docstrings and exceptions are left out for simplicity of demonstration.)
1 import sys
2
3 WHAT_TO_DEBUG = set(['io', 'core']) # change to what you need
4
5 class debug:
6 '''Decorator which helps to control what aspects of a program to debug
7 on per-function basis. Aspects are provided as list of arguments.
8 It DOESN'T slowdown functions which aren't supposed to be debugged.
9 '''
10 def __init__(self, aspects=None):
11 self.aspects = set(aspects)
12
13 def __call__(self, f):
14 if self.aspects & WHAT_TO_DEBUG:
15 def newf(*args, **kwds):
16 print >> sys.stderr, f.func_name, args, kwds
17 f_result = f(*args, **kwds)
18 print >> sys.stderr, f.func_name, "returned", f_result
19 return f_result
20 newf.__doc__ = f.__doc__
21 return newf
22 else:
23 return f
24
25 @debug(['io'])
26 def prn(x):
27 print x
28
29 @debug(['core'])
30 def mult(x, y):
31 return x * y
32
33 prn(mult(2, 2))
Easy adding methods to a class instance
Credits to John Roth.
1 class Foo:
2 def __init__(self):
3 self.x = 42
4
5 foo = Foo()
6
7 def addto(instance):
8 def decorator(f):
9 import types
10 f = types.MethodType(f, instance, instance.__class__)
11 setattr(instance, f.func_name, f)
12 return f
13 return decorator
14
15 @addto(foo)
16 def print_x(self):
17 print self.x
18
19 # foo.print_x() would print "42"
Counting function calls
1 class countcalls(object):
2 "Decorator that keeps track of the number of times a function is called."
3
4 __instances = {}
5
6 def __init__(self, f):
7 self.__f = f
8 self.__numcalls = 0
9 countcalls.__instances[f] = self
10
11 def __call__(self, *args, **kwargs):
12 self.__numcalls += 1
13 return self.__f(*args, **kwargs)
14
15 @staticmethod
16 def count(f):
17 "Return the number of times the function f was called."
18 return countcalls.__instances[f].__numcalls
19
20 @staticmethod
21 def counts():
22 "Return a dict of {function: # of calls} for all registered functions."
23 return dict([(f, countcalls.count(f)) for f in countcalls.__instances])
Alternate Counting function calls
1 class countcalls(object):
2 "Decorator that keeps track of the number of times a function is called."
3
4 __instances = {}
5
6 def __init__(self, f):
7 self.__f = f
8 self.__numcalls = 0
9 countcalls.__instances[f] = self
10
11 def __call__(self, *args, **kwargs):
12 self.__numcalls += 1
13 return self.__f(*args, **kwargs)
14
15 def count(self):
16 "Return the number of times the function f was called."
17 return countcalls.__instances[self.__f].__numcalls
18
19 @staticmethod
20 def counts():
21 "Return a dict of {function: # of calls} for all registered functions."
22 return dict([(f.__name__, countcalls.__instances[f].__numcalls) for f in countcalls.__instances])
23
24 #example
25
26 @countcalls
27 def f():
28 print 'f called'
29
30 @countcalls
31 def g():
32 print 'g called'
33
34 f()
35 f()
36 f()
37 print f.count() # prints 3
38 print countcalls.counts() # same as f.counts() or g.counts()
39 g()
40 print g.count() # prints 1
Generating Deprecation Warnings
1 import warnings
2
3 def deprecated(func):
4 '''This is a decorator which can be used to mark functions
5 as deprecated. It will result in a warning being emitted
6 when the function is used.'''
7 def new_func(*args, **kwargs):
8 warnings.warn("Call to deprecated function {}.".format(func.__name__),
9 category=DeprecationWarning)
10 return func(*args, **kwargs)
11 new_func.__name__ = func.__name__
12 new_func.__doc__ = func.__doc__
13 new_func.__dict__.update(func.__dict__)
14 return new_func
15
16 # === Examples of use ===
17
18 @deprecated
19 def some_old_function(x,y):
20 return x + y
21
22 class SomeClass:
23 @deprecated
24 def some_old_method(self, x,y):
25 return x + y
Smart deprecation warnings (with valid filenames, line numbers, etc.)
1 import warnings
2 import functools
3
4
5 def deprecated(func):
6 '''This is a decorator which can be used to mark functions
7 as deprecated. It will result in a warning being emitted
8 when the function is used.'''
9
10 @functools.wraps(func)
11 def new_func(*args, **kwargs):
12 warnings.warn_explicit(
13 "Call to deprecated function {}.".format(func.__name__),
14 category=DeprecationWarning,
15 filename=func.func_code.co_filename,
16 lineno=func.func_code.co_firstlineno + 1
17 )
18 return func(*args, **kwargs)
19 return new_func
20
21
22 ## Usage examples ##
23 @deprecated
24 def my_func():
25 pass
26
27 @other_decorators_must_be_upper
28 @deprecated
29 def my_func():
30 pass
Ignoring Deprecation Warnings
1 import warnings
2
3 def ignore_deprecation_warnings(func):
4 '''This is a decorator which can be used to ignore deprecation warnings
5 occurring in a function.'''
6 def new_func(*args, **kwargs):
7 with warnings.catch_warnings():
8 warnings.filterwarnings("ignore", category=DeprecationWarning)
9 return func(*args, **kwargs)
10 new_func.__name__ = func.__name__
11 new_func.__doc__ = func.__doc__
12 new_func.__dict__.update(func.__dict__)
13 return new_func
14
15 # === Examples of use ===
16
17 @ignore_deprecation_warnings
18 def some_function_raising_deprecation_warning():
19 warnings.warn("This is a deprecationg warning.",
20 category=DeprecationWarning)
21
22 class SomeClass:
23 @ignore_deprecation_warnings
24 def some_method_raising_deprecation_warning():
25 warnings.warn("This is a deprecationg warning.",
26 category=DeprecationWarning)
Enable/Disable Decorators
1 def unchanged(func):
2 "This decorator doesn't add any behavior"
3 return func
4
5 def disabled(func):
6 "This decorator disables the provided function, and does nothing"
7 def empty_func(*args,**kargs):
8 pass
9 return empty_func
10
11 # define this as equivalent to unchanged, for nice symmetry with disabled
12 enabled = unchanged
13
14 #
15 # Sample use
16 #
17
18 GLOBAL_ENABLE_FLAG = True
19
20 state = enabled if GLOBAL_ENABLE_FLAG else disabled
21 @state
22 def special_function_foo():
23 print "function was enabled"
Easy Dump of Function Arguments
1 def dump_args(func):
2 "This decorator dumps out the arguments passed to a function before calling it"
3 argnames = func.func_code.co_varnames[:func.func_code.co_argcount]
4 fname = func.func_name
5
6 def echo_func(*args,**kwargs):
7 print fname, ":", ', '.join(
8 '%s=%r' % entry
9 for entry in zip(argnames,args) + kwargs.items())
10 return func(*args, **kwargs)
11
12 return echo_func
13
14 @dump_args
15 def f1(a,b,c):
16 print a + b + c
17
18 f1(1, 2, 3)
Pre-/Post-Conditions
1 '''
2 Provide pre-/postconditions as function decorators.
3
4 Example usage:
5
6 >>> def in_ge20(inval):
7 ... assert inval >= 20, 'Input value < 20'
8 ...
9 >>> def out_lt30(retval, inval):
10 ... assert retval < 30, 'Return value >= 30'
11 ...
12 >>> @precondition(in_ge20)
13 ... @postcondition(out_lt30)
14 ... def inc(value):
15 ... return value + 1
16 ...
17 >>> inc(5)
18 Traceback (most recent call last):
19 ...
20 AssertionError: Input value < 20
21 >>> inc(29)
22 Traceback (most recent call last):
23 ...
24 AssertionError: Return value >= 30
25 >>> inc(20)
26 21
27
28 You can define as many pre-/postconditions for a function as you
29 like. It is also possible to specify both types of conditions at once:
30
31 >>> @conditions(in_ge20, out_lt30)
32 ... def add1(value):
33 ... return value + 1
34 ...
35 >>> add1(5)
36 Traceback (most recent call last):
37 ...
38 AssertionError: Input value < 20
39
40 An interesting feature is the ability to prevent the creation of
41 pre-/postconditions at function definition time. This makes it
42 possible to use conditions for debugging and then switch them off for
43 distribution.
44
45 >>> debug = False
46 >>> @precondition(in_ge20, debug)
47 ... def dec(value):
48 ... return value - 1
49 ...
50 >>> dec(5)
51 4
52 '''
53
54 __all__ = ['precondition', 'postcondition', 'conditions']
55
56 DEFAULT_ON = True
57
58 def precondition(precondition, use_conditions=DEFAULT_ON):
59 return conditions(precondition, None, use_conditions)
60
61 def postcondition(postcondition, use_conditions=DEFAULT_ON):
62 return conditions(None, postcondition, use_conditions)
63
64 class conditions(object):
65 __slots__ = ('__precondition', '__postcondition')
66
67 def __init__(self, pre, post, use_conditions=DEFAULT_ON):
68 if not use_conditions:
69 pre, post = None, None
70
71 self.__precondition = pre
72 self.__postcondition = post
73
74 def __call__(self, function):
75 # combine recursive wrappers (@precondition + @postcondition == @conditions)
76 pres = set((self.__precondition,))
77 posts = set((self.__postcondition,))
78
79 # unwrap function, collect distinct pre-/post conditions
80 while type(function) is FunctionWrapper:
81 pres.add(function._pre)
82 posts.add(function._post)
83 function = function._func
84
85 # filter out None conditions and build pairs of pre- and postconditions
86 conditions = map(None, filter(None, pres), filter(None, posts))
87
88 # add a wrapper for each pair (note that 'conditions' may be empty)
89 for pre, post in conditions:
90 function = FunctionWrapper(pre, post, function)
91
92 return function
93
94 class FunctionWrapper(object):
95 def __init__(self, precondition, postcondition, function):
96 self._pre = precondition
97 self._post = postcondition
98 self._func = function
99
100 def __call__(self, *args, **kwargs):
101 precondition = self._pre
102 postcondition = self._post
103
104 if precondition:
105 precondition(*args, **kwargs)
106 result = self._func(*args, **kwargs)
107 if postcondition:
108 postcondition(result, *args, **kwargs)
109 return result
110
111 def __test():
112 import doctest
113 doctest.testmod()
114
115 if __name__ == "__main__":
116 __test()
Profiling/Coverage Analysis
The code and examples are a bit longish, so I'll include a link instead: http://mg.pov.lt/blog/profiling.html
Line Tracing Individual Functions
I cobbled this together from the trace module. It allows you to decorate individual functions so their lines are traced. I think it works out to be a slightly smaller hammer than running the trace module and trying to pare back what it traces using exclusions.
1 import sys
2 import os
3 import linecache
4
5 def trace(f):
6 def globaltrace(frame, why, arg):
7 if why == "call":
8 return localtrace
9 return None
10
11 def localtrace(frame, why, arg):
12 if why == "line":
13 # record the file name and line number of every trace
14 filename = frame.f_code.co_filename
15 lineno = frame.f_lineno
16
17 bname = os.path.basename(filename)
18 print "{}({}): {}".format( bname,
19 lineno,
20 linecache.getline(filename, lineno)),
21 return localtrace
22
23 def _f(*args, **kwds):
24 sys.settrace(globaltrace)
25 result = f(*args, **kwds)
26 sys.settrace(None)
27 return result
28
29 return _f
Synchronization
Synchronize two (or more) functions on a given lock.
1 def synchronized(lock):
2 '''Synchronization decorator.'''
3
4 def wrap(f):
5 def new_function(*args, **kw):
6 lock.acquire()
7 try:
8 return f(*args, **kw)
9 finally:
10 lock.release()
11 return new_function
12 return wrap
13
14 # Example usage:
15
16 from threading import Lock
17 my_lock = Lock()
18
19 @synchronized(my_lock)
20 def critical1(*args):
21 # Interesting stuff goes here.
22 pass
23
24 @synchronized(my_lock)
25 def critical2(*args):
26 # Other interesting stuff goes here.
27 pass
Type Enforcement (accepts/returns)
Provides various degrees of type enforcement for function parameters and return values.
1 '''
2 One of three degrees of enforcement may be specified by passing
3 the 'debug' keyword argument to the decorator:
4 0 -- NONE: No type-checking. Decorators disabled.
5 #!python
6 -- MEDIUM: Print warning message to stderr. (Default)
7 2 -- STRONG: Raise TypeError with message.
8 If 'debug' is not passed to the decorator, the default level is used.
9
10 Example usage:
11 >>> NONE, MEDIUM, STRONG = 0, 1, 2
12 >>>
13 >>> @accepts(int, int, int)
14 ... @returns(float)
15 ... def average(x, y, z):
16 ... return (x + y + z) / 2
17 ...
18 >>> average(5.5, 10, 15.0)
19 TypeWarning: 'average' method accepts (int, int, int), but was given
20 (float, int, float)
21 15.25
22 >>> average(5, 10, 15)
23 TypeWarning: 'average' method returns (float), but result is (int)
24 15
25
26 Needed to cast params as floats in function def (or simply divide by 2.0).
27
28 >>> TYPE_CHECK = STRONG
29 >>> @accepts(int, debug=TYPE_CHECK)
30 ... @returns(int, debug=TYPE_CHECK)
31 ... def fib(n):
32 ... if n in (0, 1): return n
33 ... return fib(n-1) + fib(n-2)
34 ...
35 >>> fib(5.3)
36 Traceback (most recent call last):
37 ...
38 TypeError: 'fib' method accepts (int), but was given (float)
39
40 '''
41 import sys
42
43 def accepts(*types, **kw):
44 '''Function decorator. Checks decorated function's arguments are
45 of the expected types.
46
47 Parameters:
48 types -- The expected types of the inputs to the decorated function.
49 Must specify type for each parameter.
50 kw -- Optional specification of 'debug' level (this is the only valid
51 keyword argument, no other should be given).
52 debug = ( 0 | 1 | 2 )
53
54 '''
55 if not kw:
56 # default level: MEDIUM
57 debug = 1
58 else:
59 debug = kw['debug']
60 try:
61 def decorator(f):
62 def newf(*args):
63 if debug is 0:
64 return f(*args)
65 assert len(args) == len(types)
66 argtypes = tuple(map(type, args))
67 if argtypes != types:
68 msg = info(f.__name__, types, argtypes, 0)
69 if debug is 1:
70 print >> sys.stderr, 'TypeWarning: ', msg
71 elif debug is 2:
72 raise TypeError, msg
73 return f(*args)
74 newf.__name__ = f.__name__
75 return newf
76 return decorator
77 except KeyError, key:
78 raise KeyError, key + "is not a valid keyword argument"
79 except TypeError, msg:
80 raise TypeError, msg
81
82
83 def returns(ret_type, **kw):
84 '''Function decorator. Checks decorated function's return value
85 is of the expected type.
86
87 Parameters:
88 ret_type -- The expected type of the decorated function's return value.
89 Must specify type for each parameter.
90 kw -- Optional specification of 'debug' level (this is the only valid
91 keyword argument, no other should be given).
92 debug=(0 | 1 | 2)
93 '''
94 try:
95 if not kw:
96 # default level: MEDIUM
97 debug = 1
98 else:
99 debug = kw['debug']
100 def decorator(f):
101 def newf(*args):
102 result = f(*args)
103 if debug is 0:
104 return result
105 res_type = type(result)
106 if res_type != ret_type:
107 msg = info(f.__name__, (ret_type,), (res_type,), 1)
108 if debug is 1:
109 print >> sys.stderr, 'TypeWarning: ', msg
110 elif debug is 2:
111 raise TypeError, msg
112 return result
113 newf.__name__ = f.__name__
114 return newf
115 return decorator
116 except KeyError, key:
117 raise KeyError, key + "is not a valid keyword argument"
118 except TypeError, msg:
119 raise TypeError, msg
120
121 def info(fname, expected, actual, flag):
122 '''Convenience function returns nicely formatted error/warning msg.'''
123 format = lambda types: ', '.join([str(t).split("'")[1] for t in types])
124 expected, actual = format(expected), format(actual)
125 msg = "'{}' method ".format( fname )\
126 + ("accepts", "returns")[flag] + " ({}), but ".format(expected)\
127 + ("was given", "result is")[flag] + " ({})".format(actual)
128 return msg
CGI method wrapper
Handles HTML boilerplate at top and bottom of pages returned from CGI methods. Works with the cgi module. Now your request handlers can just output the interesting HTML, and let the decorator deal with all the top and bottom clutter.
(Note: the exception handler eats all exceptions, which in CGI is no big loss, since the program runs in its separate subprocess. At least here, the exception contents will be written to the output page.)
1 class CGImethod(object):
2 def __init__(self, title):
3 self.title = title
4
5 def __call__(self, fn):
6 def wrapped_fn(*args):
7 print "Content-Type: text/html\n\n"
8 print "<HTML>"
9 print "<HEAD><TITLE>{}</TITLE></HEAD>".format(self.title)
10 print "<BODY>"
11 try:
12 fn(*args)
13 except Exception, e:
14 print
15 print e
16 print
17 print "</BODY></HTML>"
18
19 return wrapped_fn
20
21 @CGImethod("Hello with Decorator")
22 def say_hello():
23 print '<h1>Hello from CGI-Land</h1>'
State Machine Implementaion
A much improved version of decorators for implementing state machines, too long to show here, is at State Machine via Decorators
This example uses Decorators to facilitate the implementation of a state machine in Python. Decorators are used to specify which methods are the event handlers for the class. In this example, actions are associated with the transitions, but it is possible with a little consideration to associate actions with states instead.
The example defines a class, MyMachine that is a state machine. Multiple instances of the class may be instantiated with each maintaining its own state. A class also may have multiple states. Here I've used gstate and tstate.
The code in the imported statedefn file gets a bit hairy, but you may not need to delve into it for your application.
1 # State Machine example Program
2
3 from statedefn import *
4
5 class MyMachine(object):
6
7 # Create Statedefn object for each state you need to keep track of.
8 # the name passed to the constructor becomes a StateVar member of the current class.
9 # i.e. if my_obj is a MyMachine object, my_obj.gstate maintains the current gstate
10 gstate = StateTable("gstate")
11 tstate = StateTable("turtle")
12
13 def __init__(self, name):
14 # must call init method of class's StateTable object. to initialize state variable
15 self.gstate.initialize(self)
16 self.tstate.initialize(self)
17 self.mname = name
18 self.a_count = 0
19 self.b_count = 0
20 self.c_count = 0
21
22 # Decorate the Event Handler virtual functions -note gstate parameter
23 @event_handler(gstate)
24 def event_a(self): pass
25
26 @event_handler(gstate)
27 def event_b(self): pass
28
29 @event_handler(gstate)
30 def event_c(self, val): pass
31
32 @event_handler(tstate)
33 def toggle(self): pass
34
35
36 # define methods to handle events.
37 def _event_a_hdlr1(self):
38 print "State 1, event A"
39 self.a_count += 1
40 def _event_b_hdlr1(self):
41 print "State 1, event B"
42 self.b_count += 1
43 def _event_c_hdlr1(self, val):
44 print "State 1, event C"
45 self.c_count += 3*val
46
47 def _event_a_hdlr2(self):
48 print "State 2, event A"
49 self.a_count += 10
50 # here we brute force the tstate to on, leave & enter functions called if state changes.
51 # turtle is object's state variable for tstate, comes from constructor argument
52 self.turtle.set_state(self, self._t_on)
53 def _event_b_hdlr2(self):
54 print "State 2, event B"
55 self.b_count += 10
56 def _event_c_hdlr2(self, val):
57 print "State 2, event C"
58 self.c_count += 2*val
59
60 def _event_a_hdlr3(self):
61 self.a_count += 100
62 print "State 3, event A"
63 def _event_b_hdlr3(self):
64 print "State 3, event B"
65 self.b_count += 100
66 # we decide here we want to go to state 2, overrrides spec in state table below.
67 # transition to next_state is made after the method exits.
68 self.gstate.next_state = self._state2
69 def _event_c_hdlr3(self, val):
70 print "State 3, event C"
71 self.c_count += 5*val
72
73 # Associate the handlers with a state. The first argument is a list of methods.
74 # One method for each event_handler decorated function of gstate. Order of methods
75 # in the list correspond to order in which the Event Handlers were declared.
76 # Second arg is the name of the state. Third argument is to be come a list of the
77 # next states.
78 # The first state created becomes the initial state.
79 _state1 = gstate.state("One", (_event_a_hdlr1, _event_b_hdlr1, _event_c_hdlr1),
80 ("Two", "Three", None))
81 _state2 = gstate.state("Two", (_event_a_hdlr2, _event_b_hdlr2, _event_c_hdlr2),
82 ("Three", None, "One"))
83 _state3 = gstate.state("Three",(_event_a_hdlr3, _event_b_hdlr3, _event_c_hdlr3),
84 (None, "One", "Two"))
85
86
87 # Declare a function that will be called when entering a new gstate.
88 # Can also declare a leave function using @on_leave_function(gstate)
89 @on_enter_function(gstate)
90 def _enter_gstate(self):
91 print "entering state ", self.gstate.name() , "of ", self.mname
92 @on_leave_function(tstate)
93 def _leave_tstate(self):
94 print "leaving state ", self.turtle.name() , "of ", self.mname
95
96
97 def _toggle_on(self):
98 print "Toggle On"
99
100 def _toggle_off(self):
101 print "Toggle Off"
102
103 _t_off = tstate.state("Off", [_toggle_on],
104 ["On"])
105 _t_on = tstate.state("On", [_toggle_off],
106 ["Off"])
107
108
109 def main():
110 big_machine = MyMachine("big")
111 lil_machine = MyMachine("lil")
112
113 big_machine.event_a()
114 lil_machine.event_a()
115 big_machine.event_a()
116 lil_machine.event_a()
117 big_machine.event_b()
118 lil_machine.event_b()
119 big_machine.event_c(4)
120 lil_machine.event_c(2)
121 big_machine.event_c(1)
122 lil_machine.event_c(3)
123 big_machine.event_b()
124 lil_machine.event_b()
125 big_machine.event_a()
126 lil_machine.event_a()
127 big_machine.event_a()
128
129 big_machine.toggle()
130 big_machine.toggle()
131 big_machine.toggle()
132
133 lil_machine.event_a()
134 big_machine.event_b()
135 lil_machine.event_b()
136 big_machine.event_c(3)
137 big_machine.event_a()
138 lil_machine.event_c(2)
139 lil_machine.event_a()
140 big_machine.event_b()
141 lil_machine.event_b()
142 big_machine.event_c(7)
143 lil_machine.event_c(1)
144
145 print "Event A count ", big_machine.a_count
146 print "Event B count ", big_machine.b_count
147 print "Event C count ", big_machine.c_count
148 print "LilMachine C count ", lil_machine.c_count
149
150 main()
And now the imported statedefn.py
1 #
2 # Support for State Machines. ref - Design Patterns by GoF
3 # Many of the methods in these classes get called behind the scenes.
4 #
5 # Notable exceptions are methods of the StateVar class.
6 #
7 # See example programs for how this module is intended to be used.
8 #
9 class StateMachineError(Exception):
10 def __init__(self, args = None):
11 self.args = args
12
13 class StateVar(object):
14 def __init__(self, initial_state):
15 self._current_state = initial_state
16 self.next_state = initial_state # publicly settable in an event handling routine.
17
18 def set_state(self, owner, new_state):
19 '''
20 Forces a state change to new_state
21 '''
22 self.next_state = new_state
23 self.__to_next_state(owner)
24
25 def __to_next_state(self, owner):
26 '''
27 The low-level state change function which calls leave state & enter state functions as
28 needed.
29
30 LeaveState and EnterState functions are called as needed when state transitions.
31 '''
32 if self.next_state is not self._current_state:
33 if hasattr(self._current_state, "leave"):
34 self._current_state.leave(owner)
35 elif hasattr(self, "leave"):
36 self.leave(owner)
37 self._current_state = self.next_state
38 if hasattr(self._current_state, "enter"):
39 self._current_state.enter(owner)
40 elif hasattr(self, "enter"):
41 self.enter(owner)
42
43 def __fctn(self, func_name):
44 '''
45 Returns the owning class's method for handling an event for the current state.
46 This method not for public consumption.
47 '''
48 vf = self._current_state.get_fe(func_name)
49 return vf
50
51 def name(self):
52 '''
53 Returns the current state name.
54 '''
55 return self._current_state.name
56
57 class STState(object):
58 def __init__(self, state_name):
59 self.name = state_name
60 self.fctn_dict = {}
61
62 def set_events(self, event_list, event_hdlr_list, next_states):
63 dictionary = self.fctn_dict
64 if not next_states:
65 def set_row(event, method):
66 dictionary[event] = [method, None]
67 map(set_row, event_list, event_hdlr_list)
68 else:
69 def set_row2(event, method, next_state):
70 dictionary[event] = [method, next_state]
71 map(set_row2, event_list, event_hdlr_list, next_states)
72 self.fctn_dict = dictionary
73
74 def get_fe(self, fctn_name):
75 return self.fctn_dict[fctn_name]
76
77 def map_next_states(self, state_dict):
78 ''' Changes second dict value from name of state to actual state.'''
79 for de in self.fctn_dict.values():
80 next_state_name = de[1]
81 if next_state_name:
82 if next_state_name in state_dict:
83 de[1] = state_dict[next_state_name]
84 else:
85 raise StateMachineError('Invalid Name for next state: {}'.format(next_state_name))
86
87
88 class StateTable(object):
89 '''
90 Magical class to define a state machine, with the help of several decorator functions
91 which follow.
92 '''
93 def __init__(self, declname):
94 self.machine_var = declname
95 self._initial_state = None
96 self._state_list = {}
97 self._event_list = []
98 self.need_initialize = 1
99
100 def initialize(self, parent):
101 '''
102 Initializes the parent class's state variable for this StateTable class.
103 Must call this method in the parent' object's __init__ method. You can have
104 Multiple state machines within a parent class. Call this method for each
105 '''
106 statevar= StateVar(self._initial_state)
107 setattr(parent, self.machine_var, statevar)
108 if hasattr(self, "enter"):
109 statevar.enter = self.enter
110 if hasattr(self, "leave"):
111 statevar.leave = self.leave
112 #Magic happens here - in the 'next state' table, translate names into state objects.
113 if self.need_initialize:
114 for xstate in list(self._state_list.values()):
115 xstate.map_next_states(self._state_list)
116 self.need_initialize = 0
117
118 def def_state(self, event_hdlr_list, name):
119 '''
120 This is used to define a state. the event handler list is a list of functions that
121 are called for corresponding events. name is the name of the state.
122 '''
123 state_table_row = STState(name)
124 if len(event_hdlr_list) != len(self._event_list):
125 raise StateMachineError('Mismatch between number of event handlers and the methods specified for the state.')
126
127 state_table_row.set_events(self._event_list, event_hdlr_list, None)
128
129 if self._initial_state is None:
130 self._initial_state = state_table_row
131 self._state_list[name] = state_table_row
132 return state_table_row
133
134 def state(self, name, event_hdlr_list, next_states):
135 state_table_row = STState(name)
136 if len(event_hdlr_list) != len(self._event_list):
137 raise StateMachineError('Mismatch between number of event handlers and the methods specified for the state.')
138 if next_states is not None and len(next_states) != len(self._event_list):
139 raise StateMachineError('Mismatch between number of event handlers and the next states specified for the state.')
140
141 state_table_row.set_events(self._event_list, event_hdlr_list, next_states)
142
143 if self._initial_state is None:
144 self._initial_state = state_table_row
145 self._state_list[name] = state_table_row
146 return state_table_row
147
148 def __add_ev_hdlr(self, func_name):
149 '''
150 Informs the class of an event handler to be added. We just need the name here. The
151 function name will later be associated with one of the functions in a list when a state is defined.
152 '''
153 self._event_list.append(func_name)
154
155 # Decorator functions ...
156 def event_handler(state_class):
157 '''
158 Declare a method that handles a type of event.
159 '''
160 def wrapper(func):
161 state_class._StateTable__add_ev_hdlr(func.__name__)
162 def obj_call(self, *args, **keywords):
163 state_var = getattr(self, state_class.machine_var)
164 funky, next_state = state_var._StateVar__fctn(func.__name__)
165 if next_state is not None:
166 state_var.next_state = next_state
167 rv = funky(self, *args, **keywords)
168 state_var._StateVar__to_next_state(self)
169 return rv
170 return obj_call
171 return wrapper
172
173 def on_enter_function(state_class):
174 '''
175 Declare that this method should be called whenever a new state is entered.
176 '''
177 def wrapper(func):
178 state_class.enter = func
179 return func
180 return wrapper
181
182 def on_leave_function(state_class):
183 '''
184 Declares that this method should be called whenever leaving a state.
185 '''
186 def wrapper(func):
187 state_class.leave = func
188 return func
189 return wrapper
C++/Java-keyword-like function decorators
@abstractMethod, @deprecatedMethod, @privateMethod, @protectedMethod, @raises, @parameterTypes, @returnType
The annotations provide run-time type checking and an alternative way to document code.
The code and documentation are long, so I offer a link: http://fightingquaker.com/pyanno/
Different Decorator Forms
There are operational differences between:
- Decorator with no arguments
- Decorator with arguments
- Decorator with wrapped class instance awareness
This example demonstrates the operational differences between the three using a skit taken from Episode 22: Bruces.
1 from sys import stdout,stderr
2 from pdb import set_trace as bp
3
4 class DecoTrace(object):
5 '''
6 Decorator class with no arguments
7
8 This can only be used for functions or methods where the instance
9 is not necessary
10
11 '''
12
13 def __init__(self, f):
14 self.f = f
15
16 def _showargs(self, *fargs, **kw):
17 print >> stderr, 'T: enter {} with args={}, kw={}'.format(self.f.__name__, str(fargs), str(kw))
18
19 def _aftercall(self, status):
20 print >> stderr, 'T: exit {} with status={}'.format(self.f.__name__, str(status))
21
22 def __call__(self, *fargs, **kw):
23 '''Pass *just* function arguments to wrapped function.'''
24 self._showargs(*fargs, **kw)
25 ret=self.f(*fargs, **kw)
26 self._aftercall(ret)
27 return ret
28
29 def __repr__(self):
30 return self.f.func_name
31
32
33 class DecoTraceWithArgs(object):
34 '''decorator class with ARGUMENTS
35
36 This can be used for unbounded functions and methods. If this wraps a
37 class instance, then extract it and pass to the wrapped method as the
38 first arg.
39 '''
40
41 def __init__(self, *dec_args, **dec_kw):
42 '''The decorator arguments are passed here. Save them for runtime.'''
43 self.dec_args = dec_args
44 self.dec_kw = dec_kw
45
46 self.label = dec_kw.get('label', 'T')
47 self.fid = dec_kw.get('stream', stderr)
48
49 def _showargs(self, *fargs, **kw):
50
51 print >> self.fid, \
52 '{}: enter {} with args={}, kw={}'.format(self.label, self.f.__name__, str(fargs), str(kw))
53 print >> self.fid, \
54 '{}: passing decorator args={}, kw={}'.format(self.label, str(self.dec_args), str(self.dec_kw))
55
56 def _aftercall(self, status):
57 print >> self.fid, '{}: exit {} with status={}'.format(self.label, self.f.__name__, str(status))
58 def _showinstance(self, instance):
59 print >> self.fid, '{}: instance={}'.format(self.label, instance)
60
61 def __call__(self, f):
62 def wrapper(*fargs, **kw):
63 '''
64 Combine decorator arguments and function arguments and pass to wrapped
65 class instance-aware function/method.
66
67 Note: the first argument cannot be "self" because we get a parse error
68 "takes at least 1 argument" unless the instance is actually included in
69 the argument list, which is redundant. If this wraps a class instance,
70 the "self" will be the first argument.
71 '''
72
73 self._showargs(*fargs, **kw)
74
75 # merge decorator keywords into the kw argument list
76 kw.update(self.dec_kw)
77
78 # Does this wrap a class instance?
79 if fargs and getattr(fargs[0], '__class__', None):
80
81 # pull out the instance and combine function and
82 # decorator args
83 instance, fargs = fargs[0], fargs[1:]+self.dec_args
84 self._showinstance(instance)
85
86 # call the method
87 ret=f(instance, *fargs, **kw)
88 else:
89 # just send in the give args and kw
90 ret=f(*(fargs + self.dec_args), **kw)
91
92 self._aftercall(ret)
93 return ret
94
95 # Save wrapped function reference
96 self.f = f
97 wrapper.__name__ = f.__name__
98 wrapper.__dict__.update(f.__dict__)
99 wrapper.__doc__ = f.__doc__
100 return wrapper
101
102
103 @DecoTrace
104 def FirstBruce(*fargs, **kwargs):
105 'Simple function using simple decorator.'
106 if fargs and fargs[0]:
107 print fargs[0]
108
109 @DecoTraceWithArgs(name="Second Bruce", standardline="G'day, Bruce!")
110 def SecondBruce(*fargs, **kwargs):
111 'Simple function using decorator with arguments.'
112 print '{}:'.format(kwargs.get('name', 'Unknown Bruce'))
113
114 if fargs and fargs[0]:
115 print fargs[0]
116 else:
117 print kwargs.get('standardline', None)
118
119 class Bruce(object):
120 'Simple class.'
121
122 def __init__(self, id):
123 self.id = id
124
125 def __str__(self):
126 return self.id
127
128 def __repr__(self):
129 return 'Bruce'
130
131 @DecoTraceWithArgs(label="Trace a class", standardline="How are yer Bruce?", stream=stdout)
132 def talk(self, *fargs, **kwargs):
133 'Simple function using decorator with arguments.'
134
135 print '{}:'.format(self)
136 if fargs and fargs[0]:
137 print fargs[0]
138 else:
139 print kwargs.get('standardline', None)
140
141 ThirdBruce = Bruce('Third Bruce')
142
143 SecondBruce()
144 FirstBruce("First Bruce: Oh, Hello Bruce!")
145 ThirdBruce.talk()
146 FirstBruce("First Bruce: Bit crook, Bruce.")
147 SecondBruce("Where's Bruce?")
148 FirstBruce("First Bruce: He's not here, Bruce")
149 ThirdBruce.talk("Blimey, s'hot in here, Bruce.")
150 FirstBruce("First Bruce: S'hot enough to boil a monkey's bum!")
151 SecondBruce("That's a strange expression, Bruce.")
152 FirstBruce("First Bruce: Well Bruce, I heard the Prime Minister use it. S'hot enough to boil a monkey's bum in 'ere, your Majesty,' he said and she smiled quietly to herself.")
153 ThirdBruce.talk("She's a good Sheila, Bruce and not at all stuck up.")
Unimplemented function replacement
Allows you to test unimplemented code in a development environment by specifying a default argument as an argument to the decorator (or you can leave it off to specify None to be returned.
1 # Annotation wrapper annotation method
2 def unimplemented(defaultval):
3 if(type(defaultval) == type(unimplemented)):
4 return lambda: None
5 else:
6 # Actual annotation
7 def unimp_wrapper(func):
8 # What we replace the function with
9 def wrapper(*arg):
10 return defaultval
11 return wrapper
12 return unimp_wrapper
Redirects stdout printing to python standard logging.
1 class LogPrinter:
2 '''LogPrinter class which serves to emulates a file object and logs
3 whatever it gets sent to a Logger object at the INFO level.'''
4 def __init__(self):
5 '''Grabs the specific logger to use for logprinting.'''
6 self.ilogger = logging.getLogger('logprinter')
7 il = self.ilogger
8 logging.basicConfig()
9 il.setLevel(logging.INFO)
10
11 def write(self, text):
12 '''Logs written output to a specific logger'''
13 self.ilogger.info(text)
14
15 def logprintinfo(func):
16 '''Wraps a method so that any calls made to print get logged instead'''
17 def pwrapper(*arg, **kwargs):
18 stdobak = sys.stdout
19 lpinstance = LogPrinter()
20 sys.stdout = lpinstance
21 try:
22 return func(*arg, **kwargs)
23 finally:
24 sys.stdout = stdobak
25 return pwrapper
Access control
This example prevents users from getting access to places where they are not authorised to go
1 class LoginCheck:
2 '''
3 This class checks whether a user
4 has logged in properly via
5 the global "check_function". If so,
6 the requested routine is called.
7 Otherwise, an alternative page is
8 displayed via the global "alt_function"
9 '''
10 def __init__(self, f):
11 self._f = f
12
13 def __call__(self, *args):
14 Status = check_function()
15 if Status is 1:
16 return self._f(*args)
17 else:
18 return alt_function()
19
20 def check_function():
21 return test
22
23 def alt_function():
24 return 'Sorry - this is the forced behaviour'
25
26 @LoginCheck
27 def display_members_page():
28 print 'This is the members page'
Example:
Events rising and handling
Please see the code and examples here: http://pypi.python.org/pypi/Decovent
Singleton
1 import functools
2
3 def singleton(cls):
4 ''' Use class as singleton. '''
5
6 cls.__new_original__ = cls.__new__
7
8 @functools.wraps(cls.__new__)
9 def singleton_new(cls, *args, **kw):
10 it = cls.__dict__.get('__it__')
11 if it is not None:
12 return it
13
14 cls.__it__ = it = cls.__new_original__(cls, *args, **kw)
15 it.__init_original__(*args, **kw)
16 return it
17
18 cls.__new__ = singleton_new
19 cls.__init_original__ = cls.__init__
20 cls.__init__ = object.__init__
21
22 return cls
23
24 #
25 # Sample use:
26 #
27
28 @singleton
29 class Foo:
30 def __new__(cls):
31 cls.x = 10
32 return object.__new__(cls)
33
34 def __init__(self):
35 assert self.x == 10
36 self.x = 15
37
38 assert Foo().x == 15
39 Foo().x = 20
40 assert Foo().x == 20
Asynchronous Call
1 from Queue import Queue
2 from threading import Thread
3
4 class asynchronous(object):
5 def __init__(self, func):
6 self.func = func
7
8 def threaded(*args, **kwargs):
9 self.queue.put(self.func(*args, **kwargs))
10
11 self.threaded = threaded
12
13 def __call__(self, *args, **kwargs):
14 return self.func(*args, **kwargs)
15
16 def start(self, *args, **kwargs):
17 self.queue = Queue()
18 thread = Thread(target=self.threaded, args=args, kwargs=kwargs);
19 thread.start();
20 return asynchronous.Result(self.queue, thread)
21
22 class NotYetDoneException(Exception):
23 def __init__(self, message):
24 self.message = message
25
26 class Result(object):
27 def __init__(self, queue, thread):
28 self.queue = queue
29 self.thread = thread
30
31 def is_done(self):
32 return not self.thread.is_alive()
33
34 def get_result(self):
35 if not self.is_done():
36 raise asynchronous.NotYetDoneException('the call has not yet completed its task')
37
38 if not hasattr(self, 'result'):
39 self.result = self.queue.get()
40
41 return self.result
42
43 if __name__ == '__main__':
44 # sample usage
45 import time
46
47 @asynchronous
48 def long_process(num):
49 time.sleep(10)
50 return num * num
51
52 result = long_process.start(12)
53
54 for i in range(20):
55 print i
56 time.sleep(1)
57
58 if result.is_done():
59 print "result {0}".format(result.get_result())
60
61
62 result2 = long_process.start(13)
63
64 try:
65 print "result2 {0}".format(result2.get_result())
66
67 except asynchronous.NotYetDoneException as ex:
68 print ex.message
Class method decorator using instance
When decorating a class method, the decorator receives an function not yet bound to an instance.
The decorator can't to do anything on the instance invocating it, unless it actually is a descriptor.
1 from functools import wraps
2
3 def decorate(f):
4 '''
5 Class method decorator specific to the instance.
6
7 It uses a descriptor to delay the definition of the
8 method wrapper.
9 '''
10 class descript(object):
11 def __init__(self, f):
12 self.f = f
13
14 def __get__(self, instance, klass):
15 if instance is None:
16 # Class method was requested
17 return self.make_unbound(klass)
18 return self.make_bound(instance)
19
20 def make_unbound(self, klass):
21 @wraps(self.f)
22 def wrapper(*args, **kwargs):
23 '''This documentation will vanish :)'''
24 raise TypeError(
25 'unbound method {}() must be called with {} instance '
26 'as first argument (got nothing instead)'.format(
27 self.f.__name__,
28 klass.__name__)
29 )
30 return wrapper
31
32 def make_bound(self, instance):
33 @wraps(self.f)
34 def wrapper(*args, **kwargs):
35 '''This documentation will disapear :)'''
36 print "Called the decorated method {} of {}".format(self.f.__name__, instance)
37 return self.f(instance, *args, **kwargs)
38 # This instance does not need the descriptor anymore,
39 # let it find the wrapper directly next time:
40 setattr(instance, self.f.__name__, wrapper)
41 return wrapper
42
43 return descript(f)
This implementation replaces the descriptor by the actual decorated function ASAP to avoid overhead, but you could keep it to do even more (counting calls, etc...)
Another Retrying Decorator
Here's another decorator for causing a function to be retried a certain number of times. This decorator is superior IMHO because it should work with any old function that raises an exception on failure.
Features:
- Works with any function that signals failure by raising an exception (I.E. just about any function)
- Supports retry delay and backoff
User can specify which exceptions are caught for retrying. E.g. networking code might be expected to raise SocketError in the event of communications difficulties, while any other exception likely indicates a bug in the code.
- Hook for custom logging
GIST: https://gist.github.com/2570004
1 #
2 # Copyright 2012 by Jeff Laughlin Consulting LLC
3 #
4 # Permission is hereby granted, free of charge, to any person obtaining a copy
5 # of this software and associated documentation files (the "Software"), to deal
6 # in the Software without restriction, including without limitation the rights
7 # to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
8 # copies of the Software, and to permit persons to whom the Software is
9 # furnished to do so, subject to the following conditions:
10 #
11 # The above copyright notice and this permission notice shall be included in
12 # all copies or substantial portions of the Software.
13 #
14 # THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
15 # IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
16 # FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
17 # AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
18 # LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
19 # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
20 # SOFTWARE.
21
22
23 import sys
24 from time import sleep
25
26
27 def example_exc_handler(tries_remaining, exception, delay):
28 """Example exception handler; prints a warning to stderr.
29
30 tries_remaining: The number of tries remaining.
31 exception: The exception instance which was raised.
32 """
33 print >> sys.stderr, "Caught '%s', %d tries remaining, sleeping for %s seconds" % (exception, tries_remaining, delay)
34
35
36 def retries(max_tries, delay=1, backoff=2, exceptions=(Exception,), hook=None):
37 """Function decorator implementing retrying logic.
38
39 delay: Sleep this many seconds * backoff * try number after failure
40 backoff: Multiply delay by this factor after each failure
41 exceptions: A tuple of exception classes; default (Exception,)
42 hook: A function with the signature myhook(tries_remaining, exception);
43 default None
44
45 The decorator will call the function up to max_tries times if it raises
46 an exception.
47
48 By default it catches instances of the Exception class and subclasses.
49 This will recover after all but the most fatal errors. You may specify a
50 custom tuple of exception classes with the 'exceptions' argument; the
51 function will only be retried if it raises one of the specified
52 exceptions.
53
54 Additionally you may specify a hook function which will be called prior
55 to retrying with the number of remaining tries and the exception instance;
56 see given example. This is primarily intended to give the opportunity to
57 log the failure. Hook is not called after failure if no retries remain.
58 """
59 def dec(func):
60 def f2(*args, **kwargs):
61 mydelay = delay
62 tries = range(max_tries)
63 tries.reverse()
64 for tries_remaining in tries:
65 try:
66 return func(*args, **kwargs)
67 except exceptions as e:
68 if tries_remaining > 0:
69 if hook is not None:
70 hook(tries_remaining, e, mydelay)
71 sleep(mydelay)
72 mydelay = mydelay * backoff
73 else:
74 raise
75 else:
76 break
77 return f2
78 return dec
Logging decorator with specified logger (or default)
This decorator will log entry and exit points of your funtion using the specified logger or it defaults to your function's module name logger.
In the current form it uses the logging.INFO level, but I can easily customized to use what ever level. Same for the entry and exit messages.
1 import functools, logging
2
3
4 log = logging.getLogger(__name__)
5 log.setLevel(logging.DEBUG)
6
7 class log_with(object):
8 '''Logging decorator that allows you to log with a
9 specific logger.
10 '''
11 # Customize these messages
12 ENTRY_MESSAGE = 'Entering {}'
13 EXIT_MESSAGE = 'Exiting {}'
14
15 def __init__(self, logger=None):
16 self.logger = logger
17
18 def __call__(self, func):
19 '''Returns a wrapper that wraps func.
20 The wrapper will log the entry and exit points of the function
21 with logging.INFO level.
22 '''
23 # set logger if it was not set earlier
24 if not self.logger:
25 logging.basicConfig()
26 self.logger = logging.getLogger(func.__module__)
27
28 @functools.wraps(func)
29 def wrapper(*args, **kwds):
30 self.logger.info(self.ENTRY_MESSAGE.format(func.__name__)) # logging level .info(). Set to .debug() if you want to
31 f_result = func(*args, **kwds)
32 self.logger.info(self.EXIT_MESSAGE.format(func.__name__)) # logging level .info(). Set to .debug() if you want to
33 return f_result
34 return wrapper
1 # Sample use and output:
2
3 if __name__ == '__main__':
4 logging.basicConfig()
5 log = logging.getLogger('custom_log')
6 log.setLevel(logging.DEBUG)
7 log.info('ciao')
8
9 @log_with(log) # user specified logger
10 def foo():
11 print 'this is foo'
12 foo()
13
14 @log_with() # using default logger
15 def foo2():
16 print 'this is foo2'
17 foo2()
1 # output
2 >>> ================================ RESTART ================================
3 >>>
4 INFO:custom_log:ciao
5 INFO:custom_log:Entering foo # uses the correct logger
6 this is foo
7 INFO:custom_log:Exiting foo
8 INFO:__main__:Entering foo2 # uses the correct logger
9 this is foo2
10 INFO:__main__:Exiting foo2
Lazy Thunkify
This decorator will cause any function to, instead of running its code, start a thread to run the code, returning a thunk (function with no args) that wait for the function's completion and returns the value (or raises the exception).
Useful if you have Computation A that takes x seconds and then uses Computation B, which takes y seconds. Instead of x+y seconds you only need max(x,y) seconds.
1 import threading, sys, functools, traceback
2
3 def lazy_thunkify(f):
4 """Make a function immediately return a function of no args which, when called,
5 waits for the result, which will start being processed in another thread."""
6
7 @functools.wraps(f)
8 def lazy_thunked(*args, **kwargs):
9 wait_event = threading.Event()
10
11 result = [None]
12 exc = [False, None]
13
14 def worker_func():
15 try:
16 func_result = f(*args, **kwargs)
17 result[0] = func_result
18 except Exception, e:
19 exc[0] = True
20 exc[1] = sys.exc_info()
21 print "Lazy thunk has thrown an exception (will be raised on thunk()):\n%s" % (
22 traceback.format_exc())
23 finally:
24 wait_event.set()
25
26 def thunk():
27 wait_event.wait()
28 if exc[0]:
29 raise exc[1][0], exc[1][1], exc[1][2]
30
31 return result[0]
32
33 threading.Thread(target=worker_func).start()
34
35 return thunk
36
37 return lazy_thunked
Example:
1 @lazy_thunkify
2 def slow_double(i):
3 print "Multiplying..."
4 time.sleep(5)
5 print "Done multiplying!"
6 return i*2
7
8
9 def maybe_multiply(x):
10 double_thunk = slow_double(x)
11 print "Thinking..."
12 time.sleep(3)
13 time.sleep(3)
14 time.sleep(1)
15 if x == 3:
16 print "Using it!"
17 res = double_thunk()
18 else:
19 print "Not using it."
20 res = None
21 return res
22
23 #both take 7 seconds
24 maybe_multiply(10)
25 maybe_multiply(3)
Aggregative decorators for generator functions
This could be a whole family of decorators. The aim is applying an aggregation function to the iterated outcome of a generator-functions.
Two interesting aggregators could be sum and average:
Examples for the two proposed decorators:
Function Timeout
Ever had a function take forever in weird edge cases? In one case, a function was extracting URIs from a long string using regular expressions, and sometimes it was running into a bug in the Python regexp engine and would take minutes rather than milliseconds. The best solution was to install a timeout using an alarm signal and simply abort processing. This can conveniently be wrapped in a decorator:
1 import signal
2 import functools
3
4 class TimeoutError(Exception): pass
5
6 def timeout(seconds, error_message = 'Function call timed out'):
7 def decorated(func):
8 def _handle_timeout(signum, frame):
9 raise TimeoutError(error_message)
10
11 def wrapper(*args, **kwargs):
12 signal.signal(signal.SIGALRM, _handle_timeout)
13 signal.alarm(seconds)
14 try:
15 result = func(*args, **kwargs)
16 finally:
17 signal.alarm(0)
18 return result
19
20 return functools.wraps(func)(wrapper)
21
22 return decorated
Example:
Collect Data Difference Caused by Decorated Function
It calls a user function to collect some data before and after the decorated function runs. To calculate difference it calls the difference calculator user function.
Example: checking page numbers of a print job: get the number of all printed pages from printer before and after the printing. Then calculate difference to get the number of pages printed by the the decorated function
1 import inspect
2 # Just in case you want to use the name of the decorator instead of difference calculator
3 # But in that case if the function decorated more than once the collected difference will be overwritten
4
5 import time
6 # Demo purposes only, the difference will be generated from time
7
8 from functools import wraps
9
10
11 def collect_data_and_calculate_difference(data_collector, difference_calculator):
12 """Returns difference of data collected before and after the decorated function,
13 plus the original return value of the decorated function. Return type: dict.
14 Keys:
15 - function name of the decorated function
16 - name of the difference calculator function
17 Values:
18 - the original return value of decorated function
19 - difference calculated by difference_calculator functions
20 Parameters: functions to collect data, and create difference from collected data
21
22 Created: 2017
23 Author: George Fischhof
24 """
25
26 current_decorator_function_name = inspect.currentframe().f_code.co_name
27 # Just in case you want to use it
28
29 def function_wrapper_because_of_parameters(decorated_function):
30 difference_calculator_name = difference_calculator.__name__
31 decorated_function_name = decorated_function.__name__
32
33 i_am_the_first_decorator = not hasattr(decorated_function, '__wrapped__')
34
35 @wraps(decorated_function)
36 def wrapper(*args, **kwargs) -> dict:
37 result_dict = dict()
38
39 before = data_collector()
40 original_result = decorated_function(*args, **kwargs)
41 after = data_collector()
42
43 my_collection = difference_calculator(before=before, after=after)
44
45 i_am_not_first_decorator_but_first_is_similar_to_me = (
46 not i_am_the_first_decorator
47 and isinstance(original_result, dict)
48 and (decorated_function_name in original_result)
49 )
50
51 if i_am_not_first_decorator_but_first_is_similar_to_me:
52 original_result[difference_calculator_name] = my_collection
53 return original_result
54 else:
55 result_dict[decorated_function_name] = original_result
56 result_dict[difference_calculator_name] = my_collection
57 return result_dict
58
59 return wrapper
60 return function_wrapper_because_of_parameters
61
62
63 # Usage
64
65
66 def collect_data_or_data_series_a():
67 time.sleep(0.5)
68 return time.time()
69
70
71 def collect_data_or_data_series_b():
72 time.sleep(0.5)
73 return time.time()
74
75
76 def calculate_difference_on_data_series_a(before, after):
77 return after - before
78
79
80 def calculate_difference_on_data_series_b(before, after):
81 return after - before
82
83
84 @collect_data_and_calculate_difference(
85 data_collector=collect_data_or_data_series_a,
86 difference_calculator=calculate_difference_on_data_series_a)
87 @collect_data_and_calculate_difference(
88 data_collector=collect_data_or_data_series_b,
89 difference_calculator=calculate_difference_on_data_series_b)
90 def do_something_that_changes_the_collected_data():
91 return 'result of decorated function...'
92
93
94 print(do_something_that_changes_the_collected_data())
95 # result dict:
96 # {'calculate_difference_on_data_series_a': 1.5010299682617188,
97 # 'do_something_that_changes_the_collected_data': 'result of decorated function...',
98 # 'calculate_difference_on_data_series_b': 0.5001623630523682}