13
13
# You should have received a copy of the GNU General Public License
14
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
17
"""A simple least-recently-used (LRU) cache."""
19
from collections import deque
21
from bzrlib import symbol_versioning
26
class _LRUNode(object):
27
"""This maintains the linked-list which is the lru internals."""
29
__slots__ = ('prev', 'next_key', 'key', 'value', 'cleanup', 'size')
31
def __init__(self, key, value, cleanup=None):
33
self.next_key = _null_key
36
self.cleanup = cleanup
37
# TODO: We could compute this 'on-the-fly' like we used to, and remove
38
# one pointer from this object, we just need to decide if it
39
# actually costs us much of anything in normal usage
46
prev_key = self.prev.key
47
return '%s(%r n:%r p:%r)' % (self.__class__.__name__, self.key,
48
self.next_key, prev_key)
50
def run_cleanup(self):
52
if self.cleanup is not None:
53
self.cleanup(self.key, self.value)
55
# cleanup might raise an exception, but we want to make sure
56
# to break refcycles, etc
24
61
class LRUCache(object):
33
70
DeprecationWarning)
34
71
after_cleanup_count = after_cleanup_size
37
self._queue = deque() # Track when things are accessed
38
self._refcount = {} # number of entries in self._queue for each key
73
# The "HEAD" of the lru linked list
74
self._most_recently_used = None
75
# The "TAIL" of the lru linked list
76
self._least_recently_used = None
39
77
self._update_max_cache(max_cache, after_cleanup_count)
41
79
def __contains__(self, key):
42
80
return key in self._cache
44
82
def __getitem__(self, key):
45
val = self._cache[key]
46
self._record_access(key)
85
# Inlined from _record_access to decrease the overhead of __getitem__
86
# We also have more knowledge about structure if __getitem__ is
87
# succeeding, then we know that self._most_recently_used must not be
89
mru = self._most_recently_used
91
# Nothing to do, this node is already at the head of the queue
93
# Remove this node from the old location
95
next_key = node.next_key
96
# benchmarking shows that the lookup of _null_key in globals is faster
97
# than the attribute lookup for (node is self._least_recently_used)
98
if next_key is _null_key:
99
# 'node' is the _least_recently_used, because it doesn't have a
100
# 'next' item. So move the current lru to the previous node.
101
self._least_recently_used = node_prev
103
node_next = cache[next_key]
104
node_next.prev = node_prev
105
node_prev.next_key = next_key
106
# Insert this node at the front of the list
107
node.next_key = mru.key
109
self._most_recently_used = node
49
113
def __len__(self):
50
114
return len(self._cache)
117
"""Walk the LRU list, only meant to be used in tests."""
118
node = self._most_recently_used
120
if node.prev is not None:
121
raise AssertionError('the _most_recently_used entry is not'
122
' supposed to have a previous entry'
124
while node is not None:
125
if node.next_key is _null_key:
126
if node is not self._least_recently_used:
127
raise AssertionError('only the last node should have'
128
' no next value: %s' % (node,))
131
node_next = self._cache[node.next_key]
132
if node_next.prev is not node:
133
raise AssertionError('inconsistency found, node.next.prev'
134
' != node: %s' % (node,))
135
if node.prev is None:
136
if node is not self._most_recently_used:
137
raise AssertionError('only the _most_recently_used should'
138
' not have a previous node: %s'
141
if node.prev.next_key != node.key:
142
raise AssertionError('inconsistency found, node.prev.next'
143
' != node: %s' % (node,))
52
147
def add(self, key, value, cleanup=None):
53
148
"""Add a new value to the cache.
55
Also, if the entry is ever removed from the queue, call cleanup.
56
Passing it the key and value being removed.
150
Also, if the entry is ever removed from the cache, call
58
153
:param key: The key to store it under
59
154
:param value: The object to store
60
155
:param cleanup: None or a function taking (key, value) to indicate
61
'value' sohuld be cleaned up.
156
'value' should be cleaned up.
159
raise ValueError('cannot use _null_key as a key')
63
160
if key in self._cache:
65
self._cache[key] = value
66
if cleanup is not None:
67
self._cleanup[key] = cleanup
68
self._record_access(key)
161
node = self._cache[key]
165
# Maintain the LRU properties, even if cleanup raises an
168
node.cleanup = cleanup
169
self._record_access(node)
171
node = _LRUNode(key, value, cleanup=cleanup)
172
self._cache[key] = node
173
self._record_access(node)
70
175
if len(self._cache) > self._max_cache:
71
176
# Trigger the cleanup
179
def cache_size(self):
180
"""Get the number of entries we will cache."""
181
return self._max_cache
74
183
def get(self, key, default=None):
75
if key in self._cache:
184
node = self._cache.get(key, None)
187
self._record_access(node)
80
191
"""Get the list of keys currently cached.
96
211
# Make sure the cache is shrunk to the correct size
97
212
while len(self._cache) > self._after_cleanup_count:
98
213
self._remove_lru()
99
# No need to compact the queue at this point, because the code that
100
# calls this would have already triggered it based on queue length
102
215
def __setitem__(self, key, value):
103
216
"""Add a value to the cache, there will be no cleanup function."""
104
217
self.add(key, value, cleanup=None)
106
def _record_access(self, key):
219
def _record_access(self, node):
107
220
"""Record that key was accessed."""
108
self._queue.append(key)
109
# Can't use setdefault because you can't += 1 the result
110
self._refcount[key] = self._refcount.get(key, 0) + 1
112
# If our access queue is too large, clean it up too
113
if len(self._queue) > self._compact_queue_length:
114
self._compact_queue()
116
def _compact_queue(self):
117
"""Compact the queue, leaving things in sorted last appended order."""
119
for item in self._queue:
120
if self._refcount[item] == 1:
121
new_queue.append(item)
123
self._refcount[item] -= 1
124
self._queue = new_queue
125
# All entries should be of the same size. There should be one entry in
126
# queue for each entry in cache, and all refcounts should == 1
127
if not (len(self._queue) == len(self._cache) ==
128
len(self._refcount) == sum(self._refcount.itervalues())):
129
raise AssertionError()
131
def _remove(self, key):
132
"""Remove an entry, making sure to maintain the invariants."""
133
cleanup = self._cleanup.pop(key, None)
134
val = self._cache.pop(key)
135
if cleanup is not None:
221
# Move 'node' to the front of the queue
222
if self._most_recently_used is None:
223
self._most_recently_used = node
224
self._least_recently_used = node
226
elif node is self._most_recently_used:
227
# Nothing to do, this node is already at the head of the queue
229
# We've taken care of the tail pointer, remove the node, and insert it
232
if node is self._least_recently_used:
233
self._least_recently_used = node.prev
234
if node.prev is not None:
235
node.prev.next_key = node.next_key
236
if node.next_key is not _null_key:
237
node_next = self._cache[node.next_key]
238
node_next.prev = node.prev
240
node.next_key = self._most_recently_used.key
241
self._most_recently_used.prev = node
242
self._most_recently_used = node
245
def _remove_node(self, node):
246
if node is self._least_recently_used:
247
self._least_recently_used = node.prev
248
self._cache.pop(node.key)
249
# If we have removed all entries, remove the head pointer as well
250
if self._least_recently_used is None:
251
self._most_recently_used = None
255
# cleanup might raise an exception, but we want to make sure to
256
# maintain the linked list
257
if node.prev is not None:
258
node.prev.next_key = node.next_key
259
if node.next_key is not _null_key:
260
node_next = self._cache[node.next_key]
261
node_next.prev = node.prev
262
# And remove this node's pointers
264
node.next_key = _null_key
139
266
def _remove_lru(self):
140
267
"""Remove one entry from the lru, and handle consequences.
200
321
self._compute_size = compute_size
201
322
if compute_size is None:
202
323
self._compute_size = len
203
# This approximates that texts are > 0.5k in size. It only really
204
# effects when we clean up the queue, so we don't want it to be too
206
324
self._update_max_size(max_size, after_cleanup_size=after_cleanup_size)
207
325
LRUCache.__init__(self, max_cache=max(int(max_size/512), 1))
209
327
def add(self, key, value, cleanup=None):
210
328
"""Add a new value to the cache.
212
Also, if the entry is ever removed from the queue, call cleanup.
213
Passing it the key and value being removed.
330
Also, if the entry is ever removed from the cache, call
215
333
:param key: The key to store it under
216
334
:param value: The object to store
217
335
:param cleanup: None or a function taking (key, value) to indicate
218
'value' sohuld be cleaned up.
336
'value' should be cleaned up.
220
if key in self._cache:
339
raise ValueError('cannot use _null_key as a key')
340
node = self._cache.get(key, None)
222
341
value_len = self._compute_size(value)
223
342
if value_len >= self._after_cleanup_size:
343
# The new value is 'too big to fit', as it would fill up/overflow
344
# the cache all by itself
345
trace.mutter('Adding the key %r to an LRUSizeCache failed.'
346
' value %d is too big to fit in a the cache'
347
' with size %d %d', key, value_len,
348
self._after_cleanup_size, self._max_size)
350
# We won't be replacing the old node, so just remove it
351
self._remove_node(node)
352
if cleanup is not None:
356
node = _LRUNode(key, value, cleanup=cleanup)
357
self._cache[key] = node
359
self._value_size -= node.size
360
node.size = value_len
225
361
self._value_size += value_len
226
self._cache[key] = value
227
if cleanup is not None:
228
self._cleanup[key] = cleanup
229
self._record_access(key)
362
self._record_access(node)
231
364
if self._value_size > self._max_size:
232
365
# Time to cleanup
added
removed
bzrlib/lru_cache.py
