Note
The source-of-truth for this package has moved to https://github.com/dart-lang/labs/tree/main/pkgs/gcloud.
The gcloud
package provides a high level idiomatic Dart interface to some of
the most widely used Google Cloud Platform services.
Currently the following services are supported:
- Cloud Datastore
- Cloud Storage
- Cloud Pub/Sub
NOTE: This package is currently experimental and published under the labs.dart.dev pub publisher in order to solicit feedback.
For packages in the labs.dart.dev publisher we generally plan to either graduate the package into a supported publisher (dart.dev, tools.dart.dev) after a period of feedback and iteration, or discontinue the package. These packages have a much higher expected rate of API and breaking changes.
Your feedback is valuable and will help us evolve this package. For general feedback, suggestions, and comments, please file an issue in the bug tracker.
The APIs in this package are all based on the generic generated APIs in the googleapis and googleapis_beta packages.
This means that the authentication model for using the APIs in this package uses the googleapis_auth package.
Note that this package is only intended for being used with the standalone VM in a server or command line application. Don't expect this package to work on the browser or in Flutter.
The code snippets below demonstrating the use of this package all assume that the following imports are present:
import 'dart:io';
import 'package:googleapis_auth/auth_io.dart' as auth;
import 'package:http/http.dart' as http;
import 'package:gcloud/db.dart';
import 'package:gcloud/storage.dart';
import 'package:gcloud/pubsub.dart';
import 'package:gcloud/service_scope.dart' as ss;
import 'package:gcloud/datastore.dart' as datastore;
The first step in using the APIs is to get an authenticated HTTP client and
with that create API class instances for accessing the different APIs. The
code below assumes that you have a Google Cloud Project called my-project
with credentials for a service account from that project stored in the file
my-project.json
.
// Read the service account credentials from the file.
var jsonCredentials = new File('my-project.json').readAsStringSync();
var credentials = new auth.ServiceAccountCredentials.fromJson(jsonCredentials);
// Get an HTTP authenticated client using the service account credentials.
var scopes = []
..addAll(datastore.Datastore.Scopes)
..addAll(Storage.SCOPES)
..addAll(PubSub.SCOPES);
var client = await auth.clientViaServiceAccount(credentials, scopes);
// Instantiate objects to access Cloud Datastore, Cloud Storage
// and Cloud Pub/Sub APIs.
var db = new DatastoreDB(
new datastore.Datastore(client, 's~my-project'));
var storage = new Storage(client, 'my-project');
var pubsub = new PubSub(client, 'my-project');
All the APIs in this package supports the use of 'service scopes'. Service scopes are described in details below.
ss.fork(() {
// register the services in the new service scope.
registerDbService(db);
registerStorageService(storage);
registerPubSubService(pubsub);
// Run application using these services.
});
The services registered with the service scope can now be reached from within all the code running in the same service scope using the below getters.
dbService.
storageService.
pubsubService.
This way it is not necessary to pass the service objects around in your code.
The gcloud
package is also integrated in the Dart appengine package. This
means the gcloud
services are available both via the appengine context and
service scopes. The authentication required to access the Google Cloud Platform
services is handled automatically.
This means that getting to the App Engine Datastore can be through either the App Engine context
var db = context.services.db;
or just using the service scope registration.
var db = dbService;
Google Cloud Datastore provide a NoSQL, schemaless database for storing non-relational data. See the product page https://cloud.google.com/datastore/ for more information.
The Cloud Datastore API provides a mapping of Dart objects to entities stored in the Datastore. The following example shows how to annotate a class to make it possible to store instances of it in the Datastore.
@db.Kind()
class Person extends db.Model {
@db.StringProperty()
String name;
@db.IntProperty()
int age;
}
The Kind
annotation tell that instances of this class can be stored. The
class must also inherit from Model
. Now to store an object into the
Datastore create an instance and use the commit
function.
var person = new Person()
..name = ''
..age = 42;
await db.commit(inserts: [person]);
The function query
is used to build a Query
object which can be run to
perform the query.
var persons = (await db.query<Person>().run()).toList();
To fetch one or multiple existing entities, use lookup
.
var key = new Person()
..name = 'UniqueName'
..age = 42
..parentKey = db.emptyKey;
var person = (await db.lookup<Person>([key])).single;
var people = await db.lookup<Person>([key1, key2]);
NOTE: This package include a lower level API provided through the class
Datastore
on top of which the DatastoreDB
API is build. The main reason
for this additional API level is to bridge the gap between the different APIs
exposed inside App Engine and through the public REST API. We reserve the
rights to modify and maybe even remove this additional layer at any time.
Google Cloud Storage provide a highly available object store (aka BLOB store). See the product page https://cloud.google.com/storage/ for more information.
In Cloud Storage the objects (BLOBs) are organized in buckets. Each bucket
has a name in a global namespace. The following code creates a new bucket
named my-bucket
and writes the content of the file my-file.txt
to the
object named my-object
.
var bucket = await storage.createBucket('my-bucket');
new File('my-file.txt').openRead().pipe(bucket.write('my-object'));
The following code will read back the object.
bucket.read('my-object').pipe(new File('my-file-copy.txt').openWrite());
Google Cloud Pub/Sub provides many-to-many, asynchronous messaging. See the product page https://cloud.google.com/pubsub/ for more information.
Cloud Pub/Sub uses two concepts for messaging. Topics are used if you want to send messages and subscriptions are used to subscribe to topics and receive the messages. This decouples the producer of a message from the consumer of a message.
The following code creates a topic and sends a simple test message:
var topic = await pubsub.createTopic('my-topic');
await topic.publishString('Hello, world!')
With the following code a subscription is created on the topic and a message is pulled using the subscription. A received message must be acknowledged when the consumer has processed it.
var subscription =
await pubsub.createSubscription('my-subscription', 'my-topic');
var pullEvent = await subscription.pull();
print(pullEvent.message.asString);
await pullEvent.acknowledge();
It is also possible to receive messages using push events instead of pulling from the subscription. To do this the subscription should be configured as a push subscription with an HTTP endpoint.
await pubsub.createSubscription(
'my-subscription',
'my-topic',
endpoint: Uri.parse('https://server.example.com/push'));
With this subscription all messages will be send to the URL provided in the
endpoint
argument. The server needs to acknowledge the reception of the
message with a 200 OK
reply.
If you want to run the end-to-end tests, a Google Cloud project is required. When running these tests the following environment variables need to be set:
GCLOUD_E2E_TEST_PROJECT
The value of the environment variable GCLOUD_E2E_TEST_PROJECT
is the name
of the Google Cloud project to use. Authentication for testing uses
Application Default Credentials locally you can provide
GOOGLE_APPLICATION_CREDENTIALS
or use
gcloud auth application-default login
.
You will also need to create indexes as follows:
gcloud --project "$GCLOUD_E2E_TEST_PROJECT" datastore indexes create test/index.yaml