Thanks for the great works dbWrapper gcloud-python , it is much easier to use cloud datastore now :)
This project focus on create an alternative work-through solution to bridge the db/ndb model and cloud datastore before official solution release, I did some experiment based on previous discussion and comment.
It is not a complete implementation yet. just a proof of concept, but it did work.
import os, sys
import google
# inject protobuf
google.__path__.append(os.path.join('lib', 'google'))
# setup clouddatastore_stub
import datastore_clouddatastore_stub
from google.appengine.api import apiproxy_stub_map
stub_map = apiproxy_stub_map.APIProxyStubMap()
dataset_id = "YOUR_DATASET_ID"
email = "YOUR_SERVICE_ACCOUNT"
key_path = "YOUR_KEY_PATH"
os.environ["APPLICATION_ID"] = dataset_id
stub = datastore_clouddatastore_stub.DatastoreCloudDatastoreStub(dataset_id, email, key_path)
stub_map.RegisterStub("datastore_v3", stub)
from google.appengine.api.memcache import memcache_stub
memcache_stub = memcache_stub.MemcacheServiceStub()
stub_map.RegisterStub("memcache", memcache_stub)
apiproxy_stub_map.apiproxy = stub_map
from google.appengine.ext import db, ndb
class AdAction(ndb.Model):
#test_int = db.IntegerProperty()
#test_string = db.TextProperty()
action_pattern = ndb.StringProperty()
advertiser = ndb.IntegerProperty()
t = AdAction.get_by_id(2)
print t
t.put()While implement the orm, there are two method come up my mind:
-
monkey patch the
datastore_rpc.py,async_get... or_make_rpc_callmethod. however, this method need to mock rpc object. I don't have a clear thought about how to do it yet. -
create a new datastore_stub and register to api_proxy, just like the
datastore_sqlite_stub.pyordatastore_file_stub.pydid in development server. This method looks like easier. Currently, I choose this approach.
The project is still under construction. Any feedback or comments are welcome.