gcloud-aio-datastore 3.0.1

Asyncio Python Client for Google Cloud Datastore

Installation

$ pip install --upgrade gcloud-aio-datastore

Usage

We’re still working on documentation; for now, this should help get you started:

from gcloud.aio.datastore import Datastore
from gcloud.aio.datastore import Key
from gcloud.aio.datastore import PathElement
from gcloud.aio.datastore import GQLQuery

ds = Datastore('my-gcloud-project', '/path/to/creds.json')
key1 = Key('my-gcloud-project', [PathElement('Kind', 'entityname')])
key2 = Key('my-gcloud-project', [PathElement('Kind', 'entityname2')])

# batched lookups
entities = await ds.lookup([key1, key2])

# convenience functions for any datastore mutations
await ds.insert(key1, {'a_boolean': True, 'meaning_of_life': 41})
await ds.update(key1, {'a_boolean': True, 'meaning_of_life': 42})
await ds.upsert(key1, {'animal': 'aardvark'})
await ds.delete(key1)

# or build your own mutation sequences with full transaction support
transaction = await ds.beginTransaction()
try:
    mutations = [
        ds.make_mutation(Operation.INSERT, key1, properties={'animal': 'sloth'}),
        ds.make_mutation(Operation.UPSERT, key1, properties={'animal': 'aardvark'}),
        ds.make_mutation(Operation.INSERT, key2, properties={'animal': 'aardvark'}),
    ]
    await ds.commit(transaction, mutations=[mutation])
except Exception:
    await ds.rollback(transaction)

# support for partial keys
partial_key = Key('my-gcloud-project', [PathElement('Kind')])
# and ID allocation or reservation
allocated_keys = await ds.allocateIds([partial_key])
await ds.reserveIds(allocated_keys)

# query support
query = GQLQuery('SELECT * FROM the_meaning_of_life WHERE answer = @answer',
                 named_bindings={'answer': 42})
results = await ds.runQuery(query, session=s)

Custom Subclasses

gcloud-aio-datastore provides class interfaces mirroring all official Google API types, ie. Key and PathElement, Entity and EntityResult, and QueryResultBatch. These types will be returned from arbitrary Datastore operations, for example Datastore.allocateIds(...) will return a list of Key entities.

For advanced usage, all of these datatypes may be overloaded. A common use-case may be to deserialize entities into more specific classes. For example, given a custom entity class such as:

class MyEntityKind(gcloud.aio.datastore.Entity):
    def __init__(self, key, properties = None) -> None:
        self.key = key
        self.is_an_aardvark = (properties or {}).get('aardvark', False)

    def __repr__(self):
        return "I'm an aardvark!" if self.is_an_aardvark else "Sorry, nope"

We can then configure gcloud-aio-datastore to serialize/deserialize from this custom entity class with:

class MyCustomDatastore(gcloud.aio.datastore.Datastore):
    entity_result_kind.entity_kind = MyEntityKind

The full list of classes which may be overridden in this way is:

class MyVeryCustomDatastore(gcloud.aio.datastore.Datastore):
    entity_result_kind = EntityResult
    entity_result_kind.entity_kind = Entity
    entity_result_kind.entity_kind.key_kind = Key
    key_kind = Key
    key_kind.path_element_kind = PathElement
    query_result_batch_kind = QueryResultBatch
    query_result_batch_kind.entity_result_kind = EntityResult

You can then drop-in the MyVeryCustomDatastore class anywhere where you previously used Datastore.

Contributing

Please see our contributing guide.