[Culprit-Finder] Seperate gae related part in cache_decorator and gitile repository to gae_libs/

appengine/findit/gae_libs/caches.py

Index: appengine/findit/gae_libs/caches.py
diff --git a/appengine/findit/gae_libs/caches.py b/appengine/findit/gae_libs/caches.py
new file mode 100644
index 0000000000000000000000000000000000000000..b1a6120dda710974c21e5d9557424a5e5bba4c07
--- /dev/null
+++ b/appengine/findit/gae_libs/caches.py
@@ -0,0 +1,75 @@
+# Copyright 2015 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+import cStringIO
+import pickle
+import zlib
+
+from google.appengine.api import memcache
+
+from lib.cache import Cache
+
+
+class PickledMemCache(Cache):
+  """A memcache-backed implementation of the interface Cache.
+
+  The data to be cached should be pickleable.
+  Limitation: size of the pickled data and key should be <= 1MB.
+  """
+  def Get(self, key):
+    return memcache.get(key)
+
+  def Set(self, key, data, expire_time=0):
+    return memcache.set(key, data, time=expire_time)
+
+
+class _CachedItemMetaData(object):
+  def __init__(self, number):
+    self.number = number
+
+
+class CompressedMemCache(Cache):
+  """A memcache-backed implementation of the interface Cache with compression.
+
+  The data to be cached would be pickled and then compressed.

[wrengr, 2016/12/06 21:52:56]

"would" -> "will"

[Sharu Jiang, 2016/12/06 23:58:16]

Done.

+  Data still > 1MB will be split into sub-piece and stored separately.
+  During retrieval, if any sub-piece is missing, None is returned.
+  """
+  CHUNK_SIZE = 990000
+
+  def Get(self, key):
+    data = memcache.get(key)
+    if isinstance(data, _CachedItemMetaData):
+      num = data.number
+      sub_keys = ['%s-%s' % (key, i) for i in range(num)]
+      all_data = memcache.get_multi(sub_keys)
+      if len(all_data) != num:  # Some data is missing.
+        return None
+
+      data_output = cStringIO.StringIO()
+      for sub_key in sub_keys:
+        data_output.write(all_data[sub_key])
+      data = data_output.getvalue()
+
+    return None if data is None else pickle.loads(zlib.decompress(data))
+
+  def Set(self, key, data, expire_time=0):
+    pickled_data = pickle.dumps(data)
+    compressed_data = zlib.compress(pickled_data)
+
+    all_data = {}
+    if len(compressed_data) > self.CHUNK_SIZE:
+      num = 0
+      for index in range(0, len(compressed_data), self.CHUNK_SIZE):

[wrengr, 2016/12/06 21:52:56]

It's better to use xrange rather than range here. No point in allocating a list
since you're just going to walk over it once.

[Sharu Jiang, 2016/12/06 23:58:16]

Done.

+        sub_key = '%s-%s' % (key, num)
+        all_data[sub_key] = compressed_data[index : index + self.CHUNK_SIZE]
+        num += 1
+
+      all_data[key] = _CachedItemMetaData(num)
+    else:
+      all_data[key] = compressed_data
+
+    keys_not_set = memcache.set_multi(all_data, time=expire_time)
+    return len(keys_not_set) == 0
+