diff options
Diffstat (limited to 'src')
-rw-r--r-- | src/cachetools/__init__.py | 596 | ||||
-rw-r--r-- | src/cachetools/cache.py | 9 | ||||
-rw-r--r-- | src/cachetools/fifo.py | 9 | ||||
-rw-r--r-- | src/cachetools/func.py | 171 | ||||
-rw-r--r-- | src/cachetools/keys.py | 52 | ||||
-rw-r--r-- | src/cachetools/lfu.py | 9 | ||||
-rw-r--r-- | src/cachetools/lru.py | 9 | ||||
-rw-r--r-- | src/cachetools/mru.py | 9 | ||||
-rw-r--r-- | src/cachetools/rr.py | 9 | ||||
-rw-r--r-- | src/cachetools/ttl.py | 9 |
10 files changed, 882 insertions, 0 deletions
diff --git a/src/cachetools/__init__.py b/src/cachetools/__init__.py new file mode 100644 index 0000000..42822f0 --- /dev/null +++ b/src/cachetools/__init__.py @@ -0,0 +1,596 @@ +"""Extensible memoizing collections and decorators.""" + +__all__ = ( + "Cache", + "FIFOCache", + "LFUCache", + "LRUCache", + "MRUCache", + "RRCache", + "TTLCache", + "cached", + "cachedmethod", +) + +__version__ = "4.2.4" + +import collections +import collections.abc +import functools +import random +import time + +from .keys import hashkey + + +class _DefaultSize: + + __slots__ = () + + def __getitem__(self, _): + return 1 + + def __setitem__(self, _, value): + assert value == 1 + + def pop(self, _): + return 1 + + +class Cache(collections.abc.MutableMapping): + """Mutable mapping to serve as a simple cache or cache base class.""" + + __marker = object() + + __size = _DefaultSize() + + def __init__(self, maxsize, getsizeof=None): + if getsizeof: + self.getsizeof = getsizeof + if self.getsizeof is not Cache.getsizeof: + self.__size = dict() + self.__data = dict() + self.__currsize = 0 + self.__maxsize = maxsize + + def __repr__(self): + return "%s(%r, maxsize=%r, currsize=%r)" % ( + self.__class__.__name__, + list(self.__data.items()), + self.__maxsize, + self.__currsize, + ) + + def __getitem__(self, key): + try: + return self.__data[key] + except KeyError: + return self.__missing__(key) + + def __setitem__(self, key, value): + maxsize = self.__maxsize + size = self.getsizeof(value) + if size > maxsize: + raise ValueError("value too large") + if key not in self.__data or self.__size[key] < size: + while self.__currsize + size > maxsize: + self.popitem() + if key in self.__data: + diffsize = size - self.__size[key] + else: + diffsize = size + self.__data[key] = value + self.__size[key] = size + self.__currsize += diffsize + + def __delitem__(self, key): + size = self.__size.pop(key) + del self.__data[key] + self.__currsize -= size + + def __contains__(self, key): + return key in self.__data + + def __missing__(self, key): + raise KeyError(key) + + def __iter__(self): + return iter(self.__data) + + def __len__(self): + return len(self.__data) + + def get(self, key, default=None): + if key in self: + return self[key] + else: + return default + + def pop(self, key, default=__marker): + if key in self: + value = self[key] + del self[key] + elif default is self.__marker: + raise KeyError(key) + else: + value = default + return value + + def setdefault(self, key, default=None): + if key in self: + value = self[key] + else: + self[key] = value = default + return value + + @property + def maxsize(self): + """The maximum size of the cache.""" + return self.__maxsize + + @property + def currsize(self): + """The current size of the cache.""" + return self.__currsize + + @staticmethod + def getsizeof(value): + """Return the size of a cache element's value.""" + return 1 + + +class FIFOCache(Cache): + """First In First Out (FIFO) cache implementation.""" + + def __init__(self, maxsize, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__order = collections.OrderedDict() + + def __setitem__(self, key, value, cache_setitem=Cache.__setitem__): + cache_setitem(self, key, value) + try: + self.__order.move_to_end(key) + except KeyError: + self.__order[key] = None + + def __delitem__(self, key, cache_delitem=Cache.__delitem__): + cache_delitem(self, key) + del self.__order[key] + + def popitem(self): + """Remove and return the `(key, value)` pair first inserted.""" + try: + key = next(iter(self.__order)) + except StopIteration: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + +class LFUCache(Cache): + """Least Frequently Used (LFU) cache implementation.""" + + def __init__(self, maxsize, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__counter = collections.Counter() + + def __getitem__(self, key, cache_getitem=Cache.__getitem__): + value = cache_getitem(self, key) + if key in self: # __missing__ may not store item + self.__counter[key] -= 1 + return value + + def __setitem__(self, key, value, cache_setitem=Cache.__setitem__): + cache_setitem(self, key, value) + self.__counter[key] -= 1 + + def __delitem__(self, key, cache_delitem=Cache.__delitem__): + cache_delitem(self, key) + del self.__counter[key] + + def popitem(self): + """Remove and return the `(key, value)` pair least frequently used.""" + try: + ((key, _),) = self.__counter.most_common(1) + except ValueError: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + +class LRUCache(Cache): + """Least Recently Used (LRU) cache implementation.""" + + def __init__(self, maxsize, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__order = collections.OrderedDict() + + def __getitem__(self, key, cache_getitem=Cache.__getitem__): + value = cache_getitem(self, key) + if key in self: # __missing__ may not store item + self.__update(key) + return value + + def __setitem__(self, key, value, cache_setitem=Cache.__setitem__): + cache_setitem(self, key, value) + self.__update(key) + + def __delitem__(self, key, cache_delitem=Cache.__delitem__): + cache_delitem(self, key) + del self.__order[key] + + def popitem(self): + """Remove and return the `(key, value)` pair least recently used.""" + try: + key = next(iter(self.__order)) + except StopIteration: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + def __update(self, key): + try: + self.__order.move_to_end(key) + except KeyError: + self.__order[key] = None + + +class MRUCache(Cache): + """Most Recently Used (MRU) cache implementation.""" + + def __init__(self, maxsize, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__order = collections.OrderedDict() + + def __getitem__(self, key, cache_getitem=Cache.__getitem__): + value = cache_getitem(self, key) + if key in self: # __missing__ may not store item + self.__update(key) + return value + + def __setitem__(self, key, value, cache_setitem=Cache.__setitem__): + cache_setitem(self, key, value) + self.__update(key) + + def __delitem__(self, key, cache_delitem=Cache.__delitem__): + cache_delitem(self, key) + del self.__order[key] + + def popitem(self): + """Remove and return the `(key, value)` pair most recently used.""" + try: + key = next(iter(self.__order)) + except StopIteration: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + def __update(self, key): + try: + self.__order.move_to_end(key, last=False) + except KeyError: + self.__order[key] = None + + +class RRCache(Cache): + """Random Replacement (RR) cache implementation.""" + + def __init__(self, maxsize, choice=random.choice, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__choice = choice + + @property + def choice(self): + """The `choice` function used by the cache.""" + return self.__choice + + def popitem(self): + """Remove and return a random `(key, value)` pair.""" + try: + key = self.__choice(list(self)) + except IndexError: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + +class _Timer: + def __init__(self, timer): + self.__timer = timer + self.__nesting = 0 + + def __call__(self): + if self.__nesting == 0: + return self.__timer() + else: + return self.__time + + def __enter__(self): + if self.__nesting == 0: + self.__time = time = self.__timer() + else: + time = self.__time + self.__nesting += 1 + return time + + def __exit__(self, *exc): + self.__nesting -= 1 + + def __reduce__(self): + return _Timer, (self.__timer,) + + def __getattr__(self, name): + return getattr(self.__timer, name) + + +class _Link: + + __slots__ = ("key", "expire", "next", "prev") + + def __init__(self, key=None, expire=None): + self.key = key + self.expire = expire + + def __reduce__(self): + return _Link, (self.key, self.expire) + + def unlink(self): + next = self.next + prev = self.prev + prev.next = next + next.prev = prev + + +class TTLCache(Cache): + """LRU Cache implementation with per-item time-to-live (TTL) value.""" + + def __init__(self, maxsize, ttl, timer=time.monotonic, getsizeof=None): + Cache.__init__(self, maxsize, getsizeof) + self.__root = root = _Link() + root.prev = root.next = root + self.__links = collections.OrderedDict() + self.__timer = _Timer(timer) + self.__ttl = ttl + + def __contains__(self, key): + try: + link = self.__links[key] # no reordering + except KeyError: + return False + else: + return not (link.expire < self.__timer()) + + def __getitem__(self, key, cache_getitem=Cache.__getitem__): + try: + link = self.__getlink(key) + except KeyError: + expired = False + else: + expired = link.expire < self.__timer() + if expired: + return self.__missing__(key) + else: + return cache_getitem(self, key) + + def __setitem__(self, key, value, cache_setitem=Cache.__setitem__): + with self.__timer as time: + self.expire(time) + cache_setitem(self, key, value) + try: + link = self.__getlink(key) + except KeyError: + self.__links[key] = link = _Link(key) + else: + link.unlink() + link.expire = time + self.__ttl + link.next = root = self.__root + link.prev = prev = root.prev + prev.next = root.prev = link + + def __delitem__(self, key, cache_delitem=Cache.__delitem__): + cache_delitem(self, key) + link = self.__links.pop(key) + link.unlink() + if link.expire < self.__timer(): + raise KeyError(key) + + def __iter__(self): + root = self.__root + curr = root.next + while curr is not root: + # "freeze" time for iterator access + with self.__timer as time: + if not (curr.expire < time): + yield curr.key + curr = curr.next + + def __len__(self): + root = self.__root + curr = root.next + time = self.__timer() + count = len(self.__links) + while curr is not root and curr.expire < time: + count -= 1 + curr = curr.next + return count + + def __setstate__(self, state): + self.__dict__.update(state) + root = self.__root + root.prev = root.next = root + for link in sorted(self.__links.values(), key=lambda obj: obj.expire): + link.next = root + link.prev = prev = root.prev + prev.next = root.prev = link + self.expire(self.__timer()) + + def __repr__(self, cache_repr=Cache.__repr__): + with self.__timer as time: + self.expire(time) + return cache_repr(self) + + @property + def currsize(self): + with self.__timer as time: + self.expire(time) + return super().currsize + + @property + def timer(self): + """The timer function used by the cache.""" + return self.__timer + + @property + def ttl(self): + """The time-to-live value of the cache's items.""" + return self.__ttl + + def expire(self, time=None): + """Remove expired items from the cache.""" + if time is None: + time = self.__timer() + root = self.__root + curr = root.next + links = self.__links + cache_delitem = Cache.__delitem__ + while curr is not root and curr.expire < time: + cache_delitem(self, curr.key) + del links[curr.key] + next = curr.next + curr.unlink() + curr = next + + def clear(self): + with self.__timer as time: + self.expire(time) + Cache.clear(self) + + def get(self, *args, **kwargs): + with self.__timer: + return Cache.get(self, *args, **kwargs) + + def pop(self, *args, **kwargs): + with self.__timer: + return Cache.pop(self, *args, **kwargs) + + def setdefault(self, *args, **kwargs): + with self.__timer: + return Cache.setdefault(self, *args, **kwargs) + + def popitem(self): + """Remove and return the `(key, value)` pair least recently used that + has not already expired. + + """ + with self.__timer as time: + self.expire(time) + try: + key = next(iter(self.__links)) + except StopIteration: + raise KeyError("%s is empty" % type(self).__name__) from None + else: + return (key, self.pop(key)) + + def __getlink(self, key): + value = self.__links[key] + self.__links.move_to_end(key) + return value + + +def cached(cache, key=hashkey, lock=None): + """Decorator to wrap a function with a memoizing callable that saves + results in a cache. + + """ + + def decorator(func): + if cache is None: + + def wrapper(*args, **kwargs): + return func(*args, **kwargs) + + elif lock is None: + + def wrapper(*args, **kwargs): + k = key(*args, **kwargs) + try: + return cache[k] + except KeyError: + pass # key not found + v = func(*args, **kwargs) + try: + cache[k] = v + except ValueError: + pass # value too large + return v + + else: + + def wrapper(*args, **kwargs): + k = key(*args, **kwargs) + try: + with lock: + return cache[k] + except KeyError: + pass # key not found + v = func(*args, **kwargs) + # in case of a race, prefer the item already in the cache + try: + with lock: + return cache.setdefault(k, v) + except ValueError: + return v # value too large + + return functools.update_wrapper(wrapper, func) + + return decorator + + +def cachedmethod(cache, key=hashkey, lock=None): + """Decorator to wrap a class or instance method with a memoizing + callable that saves results in a cache. + + """ + + def decorator(method): + if lock is None: + + def wrapper(self, *args, **kwargs): + c = cache(self) + if c is None: + return method(self, *args, **kwargs) + k = key(*args, **kwargs) + try: + return c[k] + except KeyError: + pass # key not found + v = method(self, *args, **kwargs) + try: + c[k] = v + except ValueError: + pass # value too large + return v + + else: + + def wrapper(self, *args, **kwargs): + c = cache(self) + if c is None: + return method(self, *args, **kwargs) + k = key(*args, **kwargs) + try: + with lock(self): + return c[k] + except KeyError: + pass # key not found + v = method(self, *args, **kwargs) + # in case of a race, prefer the item already in the cache + try: + with lock(self): + return c.setdefault(k, v) + except ValueError: + return v # value too large + + return functools.update_wrapper(wrapper, method) + + return decorator diff --git a/src/cachetools/cache.py b/src/cachetools/cache.py new file mode 100644 index 0000000..ee5269d --- /dev/null +++ b/src/cachetools/cache.py @@ -0,0 +1,9 @@ +import warnings + +from . import Cache + +warnings.warn( + "cachetools.cache is deprecated, please use cachetools.Cache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/fifo.py b/src/cachetools/fifo.py new file mode 100644 index 0000000..a29b789 --- /dev/null +++ b/src/cachetools/fifo.py @@ -0,0 +1,9 @@ +import warnings + +from . import FIFOCache + +warnings.warn( + "cachetools.fifo is deprecated, please use cachetools.FIFOCache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/func.py b/src/cachetools/func.py new file mode 100644 index 0000000..01702c2 --- /dev/null +++ b/src/cachetools/func.py @@ -0,0 +1,171 @@ +"""`functools.lru_cache` compatible memoizing function decorators.""" + +__all__ = ("fifo_cache", "lfu_cache", "lru_cache", "mru_cache", "rr_cache", "ttl_cache") + +import collections +import functools +import math +import random +import time + +try: + from threading import RLock +except ImportError: # pragma: no cover + from dummy_threading import RLock + +from . import FIFOCache, LFUCache, LRUCache, MRUCache, RRCache, TTLCache +from . import keys + + +_CacheInfo = collections.namedtuple( + "CacheInfo", ["hits", "misses", "maxsize", "currsize"] +) + + +class _UnboundCache(dict): + @property + def maxsize(self): + return None + + @property + def currsize(self): + return len(self) + + +class _UnboundTTLCache(TTLCache): + def __init__(self, ttl, timer): + TTLCache.__init__(self, math.inf, ttl, timer) + + @property + def maxsize(self): + return None + + +def _cache(cache, typed): + maxsize = cache.maxsize + + def decorator(func): + key = keys.typedkey if typed else keys.hashkey + lock = RLock() + stats = [0, 0] + + def wrapper(*args, **kwargs): + k = key(*args, **kwargs) + with lock: + try: + v = cache[k] + stats[0] += 1 + return v + except KeyError: + stats[1] += 1 + v = func(*args, **kwargs) + # in case of a race, prefer the item already in the cache + try: + with lock: + return cache.setdefault(k, v) + except ValueError: + return v # value too large + + def cache_info(): + with lock: + hits, misses = stats + maxsize = cache.maxsize + currsize = cache.currsize + return _CacheInfo(hits, misses, maxsize, currsize) + + def cache_clear(): + with lock: + try: + cache.clear() + finally: + stats[:] = [0, 0] + + wrapper.cache_info = cache_info + wrapper.cache_clear = cache_clear + wrapper.cache_parameters = lambda: {"maxsize": maxsize, "typed": typed} + functools.update_wrapper(wrapper, func) + return wrapper + + return decorator + + +def fifo_cache(maxsize=128, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a First In First Out (FIFO) + algorithm. + + """ + if maxsize is None: + return _cache(_UnboundCache(), typed) + elif callable(maxsize): + return _cache(FIFOCache(128), typed)(maxsize) + else: + return _cache(FIFOCache(maxsize), typed) + + +def lfu_cache(maxsize=128, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a Least Frequently Used (LFU) + algorithm. + + """ + if maxsize is None: + return _cache(_UnboundCache(), typed) + elif callable(maxsize): + return _cache(LFUCache(128), typed)(maxsize) + else: + return _cache(LFUCache(maxsize), typed) + + +def lru_cache(maxsize=128, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a Least Recently Used (LRU) + algorithm. + + """ + if maxsize is None: + return _cache(_UnboundCache(), typed) + elif callable(maxsize): + return _cache(LRUCache(128), typed)(maxsize) + else: + return _cache(LRUCache(maxsize), typed) + + +def mru_cache(maxsize=128, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a Most Recently Used (MRU) + algorithm. + """ + if maxsize is None: + return _cache(_UnboundCache(), typed) + elif callable(maxsize): + return _cache(MRUCache(128), typed)(maxsize) + else: + return _cache(MRUCache(maxsize), typed) + + +def rr_cache(maxsize=128, choice=random.choice, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a Random Replacement (RR) + algorithm. + + """ + if maxsize is None: + return _cache(_UnboundCache(), typed) + elif callable(maxsize): + return _cache(RRCache(128, choice), typed)(maxsize) + else: + return _cache(RRCache(maxsize, choice), typed) + + +def ttl_cache(maxsize=128, ttl=600, timer=time.monotonic, typed=False): + """Decorator to wrap a function with a memoizing callable that saves + up to `maxsize` results based on a Least Recently Used (LRU) + algorithm with a per-item time-to-live (TTL) value. + """ + if maxsize is None: + return _cache(_UnboundTTLCache(ttl, timer), typed) + elif callable(maxsize): + return _cache(TTLCache(128, ttl, timer), typed)(maxsize) + else: + return _cache(TTLCache(maxsize, ttl, timer), typed) diff --git a/src/cachetools/keys.py b/src/cachetools/keys.py new file mode 100644 index 0000000..13630a4 --- /dev/null +++ b/src/cachetools/keys.py @@ -0,0 +1,52 @@ +"""Key functions for memoizing decorators.""" + +__all__ = ("hashkey", "typedkey") + + +class _HashedTuple(tuple): + """A tuple that ensures that hash() will be called no more than once + per element, since cache decorators will hash the key multiple + times on a cache miss. See also _HashedSeq in the standard + library functools implementation. + + """ + + __hashvalue = None + + def __hash__(self, hash=tuple.__hash__): + hashvalue = self.__hashvalue + if hashvalue is None: + self.__hashvalue = hashvalue = hash(self) + return hashvalue + + def __add__(self, other, add=tuple.__add__): + return _HashedTuple(add(self, other)) + + def __radd__(self, other, add=tuple.__add__): + return _HashedTuple(add(other, self)) + + def __getstate__(self): + return {} + + +# used for separating keyword arguments; we do not use an object +# instance here so identity is preserved when pickling/unpickling +_kwmark = (_HashedTuple,) + + +def hashkey(*args, **kwargs): + """Return a cache key for the specified hashable arguments.""" + + if kwargs: + return _HashedTuple(args + sum(sorted(kwargs.items()), _kwmark)) + else: + return _HashedTuple(args) + + +def typedkey(*args, **kwargs): + """Return a typed cache key for the specified hashable arguments.""" + + key = hashkey(*args, **kwargs) + key += tuple(type(v) for v in args) + key += tuple(type(v) for _, v in sorted(kwargs.items())) + return key diff --git a/src/cachetools/lfu.py b/src/cachetools/lfu.py new file mode 100644 index 0000000..5c9acef --- /dev/null +++ b/src/cachetools/lfu.py @@ -0,0 +1,9 @@ +import warnings + +from . import LFUCache + +warnings.warn( + "cachetools.lfu is deprecated, please use cachetools.LFUCache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/lru.py b/src/cachetools/lru.py new file mode 100644 index 0000000..48ddb36 --- /dev/null +++ b/src/cachetools/lru.py @@ -0,0 +1,9 @@ +import warnings + +from . import LRUCache + +warnings.warn( + "cachetools.lru is deprecated, please use cachetools.LRUCache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/mru.py b/src/cachetools/mru.py new file mode 100644 index 0000000..a486dc4 --- /dev/null +++ b/src/cachetools/mru.py @@ -0,0 +1,9 @@ +import warnings + +from . import MRUCache + +warnings.warn( + "cachetools.mru is deprecated, please use cachetools.MRUCache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/rr.py b/src/cachetools/rr.py new file mode 100644 index 0000000..81331bc --- /dev/null +++ b/src/cachetools/rr.py @@ -0,0 +1,9 @@ +import warnings + +from . import RRCache + +warnings.warn( + "cachetools.rr is deprecated, please use cachetools.RRCache", + DeprecationWarning, + stacklevel=2, +) diff --git a/src/cachetools/ttl.py b/src/cachetools/ttl.py new file mode 100644 index 0000000..ef343da --- /dev/null +++ b/src/cachetools/ttl.py @@ -0,0 +1,9 @@ +import warnings + +from . import TTLCache + +warnings.warn( + "cachetools.ttl is deprecated, please use cachetools.TTLCache", + DeprecationWarning, + stacklevel=2, +) |