gcloud-rest-datastore 6.0.0

Python Client for Google Cloud Datastore

This is a shared codebase for gcloud-rest-datastore and gcloud-rest-datastore

Installation

$ pip install --upgrade gcloud-{aio,rest}-datastore

Usage

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

from gcloud.rest.datastore import Datastore
from gcloud.rest.datastore import Direction
from gcloud.rest.datastore import Filter
from gcloud.rest.datastore import GQLQuery
from gcloud.rest.datastore import Key
from gcloud.rest.datastore import PathElement
from gcloud.rest.datastore import PropertyFilter
from gcloud.rest.datastore import PropertyFilterOperator
from gcloud.rest.datastore import PropertyOrder
from gcloud.rest.datastore import Query
from gcloud.rest.datastore import Value

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
property_filter = PropertyFilter(prop='answer',
                                 operator=PropertyFilterOperator.EQUAL,
                                 value=Value(42))
property_order = PropertyOrder(prop='length',
                               direction=Direction.DESCENDING)
query = Query(kind='the_meaning_of_life',
              query_filter=Filter(property_filter),
              order=property_order)
results = await ds.runQuery(query, session=s)

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

# close the HTTP session
# Note that other options include:
# * providing your own session: `Datastore(.., session=session)`
# * using a context manager: `async with Datastore(..) as ds:`
await ds.close()

Custom Subclasses

gcloud-rest-datastore provides class interfaces mirroring all official Google API types, ie. Key and PathElement, Entity and EntityResult, QueryResultBatch, and Value. 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.rest.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-rest-datastore to serialize/deserialize from this custom entity class with:

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

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

class MyVeryCustomDatastore(gcloud.rest.datastore.Datastore):
    datastore_operation_kind = DatastoreOperation
    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
    value_kind = Value
    value_kind.key_kind = Key

class MyVeryCustomQuery(gcloud.rest.datastore.Query):
    value_kind = Value

class MyVeryCustomGQLQuery(gcloud.rest.datastore.GQLQuery):
    value_kind = Value

You can then drop-in the MyVeryCustomDatastore class anywhere where you previously used Datastore and do the same for Query and GQLQuery.

To override any sub-key, you’ll need to override any parents which use it. For example, if you want to use a custom Key kind and be able to use queries with it, you will need to implement your own Value, Query, and GQLQuery classes and wire them up to the rest of the custom classes:

class MyKey(gcloud.rest.datastore.Key):
    pass

class MyValue(gcloud.rest.datastore.Value):
    key_kind = MyKey

class MyEntity(gcloud.rest.datastore.Entity):
    key_kind = MyKey
    value_kind = MyValue

class MyEntityResult(gcloud.rest.datastore.EntityResult):
    entity_kind = MyEntity

class MyQueryResultBatch(gcloud.rest.datastore.QueryResultBatch):
    entity_result_kind = MyEntityResult

class MyDatastore(gcloud.rest.datastore.Datastore):
    key_kind = MyKey
    entity_result_kind = MyEntityResult
    query_result_batch = MyQueryResultBatch
    value_kind = MyValue

class MyQuery(gcloud.rest.datastore.Query):
    value_kind = MyValue

class MyGQLQuery(gcloud.rest.datastore.GQLQuery):
    value_kind = MyValue

Contributing

Please see our contributing guide.