[feat] full caching for each sync_id session

The CouchDB backend implementation was accessing CouchDB too many times
for the same values. Those values are known inside the same sync_id,
which is the id of current sync session.
This commit adds caching for all redundant calls to Couch inside the
same sync_id for each replica.
Refactoring is still needed, but for now couch.py works normally as if
caching is not present, while sync.py injects the cache as a attribute
to enable it. This needs a simpler implementation.

3 files changed, 27 insertions, 7 deletions

diff --git a/common/src/leap/soledad/common/couch.py b/common/src/leap/soledad/common/couch.py
index 3b120cd1..8c152353 100644
--- a/common/src/leap/soledad/common/couch.py
+++ b/common/src/leap/soledad/common/couch.py
@@ -434,7 +434,14 @@ class CouchDatabase(CommonBackend):
             self._set_replica_uid(replica_uid)
         if ensure_ddocs:
             self.ensure_ddocs_on_db()
-        self.cache = {}
+        self._cache = None
+
+    @property
+    def cache(self):
+        if self._cache is not None:
+            return self._cache
+        else:
+            return {}
 
     def ensure_ddocs_on_db(self):
         """
@@ -513,7 +520,10 @@ class CouchDatabase(CommonBackend):
         :rtype: str
         """
         if self._real_replica_uid is not None:
+            self.cache[self._url] = {'replica_uid': self._real_replica_uid}
             return self._real_replica_uid
+        if self._url in self.cache:
+            return self.cache[self._url]['replica_uid']
         try:
             # grab replica_uid from server
             doc = self._database['u1db_config']
@@ -551,10 +561,13 @@ class CouchDatabase(CommonBackend):
                                              unknown reason.
         """
         # query a couch list function
+        if self.replica_uid + '_gen' in self.cache:
+            return self.cache[self.replica_uid + '_gen']['generation']
         ddoc_path = ['_design', 'transactions', '_list', 'generation', 'log']
         res = self._database.resource(*ddoc_path)
         try:
             response = res.get_json()
+            self.cache[self.replica_uid + '_gen'] = response[2]
             return response[2]['generation']
         except ResourceNotFound as e:
             raise_missing_design_doc_error(e, ddoc_path)
@@ -582,11 +595,15 @@ class CouchDatabase(CommonBackend):
                                              design document for an yet
                                              unknown reason.
         """
+        if self.replica_uid + '_gen' in self.cache:
+            response = self.cache[self.replica_uid + '_gen']
+            return (response['generation'], response['transaction_id'])
         # query a couch list function
         ddoc_path = ['_design', 'transactions', '_list', 'generation', 'log']
         res = self._database.resource(*ddoc_path)
         try:
             response = res.get_json()
+            self.cache[self.replica_uid + '_gen'] = response[2]
             return (response[2]['generation'], response[2]['transaction_id'])
         except ResourceNotFound as e:
             raise_missing_design_doc_error(e, ddoc_path)
@@ -846,6 +863,10 @@ class CouchDatabase(CommonBackend):
                 doc.doc_id, body=buf.getvalue(), headers=envelope.headers)
         except ResourceConflict:
             raise RevisionConflict()
+        if self.replica_uid + '_gen' in self.cache:
+            gen_info = self.cache[self.replica_uid + '_gen']
+            gen_info['generation'] += 1
+            gen_info['transaction_id'] = transactions[-1][1]
 
     def put_doc(self, doc):
         """
@@ -1366,7 +1387,7 @@ class CouchServerState(ServerState):
     Inteface of the WSGI server with the CouchDB backend.
     """
 
-    def __init__(self, couch_url, cache=None):
+    def __init__(self, couch_url):
         """
         Initialize the couch server state.
 
@@ -1374,7 +1395,6 @@ class CouchServerState(ServerState):
         :type couch_url: str
         """
         self.couch_url = couch_url
-        self.cache = cache or {}
 
     def open_database(self, dbname):
         """
@@ -1390,7 +1410,6 @@ class CouchServerState(ServerState):
             self.couch_url,
             dbname,
             ensure_ddocs=False)
-        db.cache = self.cache
         return db
 
     def ensure_database(self, dbname):
diff --git a/server/src/leap/soledad/server/__init__.py b/server/src/leap/soledad/server/__init__.py
index 58f2b0b1..1b795016 100644
--- a/server/src/leap/soledad/server/__init__.py
+++ b/server/src/leap/soledad/server/__init__.py
@@ -111,7 +111,6 @@ from leap.soledad.server.sync import (
 
 from leap.soledad.common import SHARED_DB_NAME
 from leap.soledad.common.couch import CouchServerState
-from leap.soledad.server.caching import get_cache_for
 
 old_tsafe = tsafe
 
@@ -304,8 +303,7 @@ def load_configuration(file_path):
 
 def application(environ, start_response):
     conf = load_configuration('/etc/leap/soledad-server.conf')
-    cache = get_cache_for('replica_cache')
-    state = CouchServerState(conf['couch_url'], cache=cache)
+    state = CouchServerState(conf['couch_url'])
     # WSGI application that may be used by `twistd -web`
     application = GzipMiddleware(
         SoledadTokenAuthMiddleware(SoledadApp(state)))
diff --git a/server/src/leap/soledad/server/sync.py b/server/src/leap/soledad/server/sync.py
index 64f7e4e7..262b6769 100644
--- a/server/src/leap/soledad/server/sync.py
+++ b/server/src/leap/soledad/server/sync.py
@@ -20,6 +20,7 @@ Server side synchronization infrastructure.
 from u1db import sync, Document
 from u1db.remote import http_app
 from leap.soledad.server.state import ServerSyncState
+from leap.soledad.server.caching import get_cache_for
 
 
 MAX_REQUEST_SIZE = 200  # in Mb
@@ -188,6 +189,8 @@ class SyncResource(http_app.SyncResource):
             db, self.replica_uid = self.state.ensure_database(self.dbname)
         else:
             db = self.state.open_database(self.dbname)
+        cache = get_cache_for('db-' + sync_id + db.replica_uid)
+        db._cache = cache
         # validate the information the client has about server replica
         db.validate_gen_and_trans_id(
             last_known_generation, last_known_trans_id)