Manage Google Cloud Storage assets

Operations on GCS assets (connections, buckets, objects).

In general, these should be:

• Created in top-down order (connection, then bucket, then object)
• Deleted in bottom-up order (objects, then buckets, then connections)¹

Asset structure

Connection

A Google Cloud Storage connection requires a name and qualifiedName. For creation, specific settings are also required to distinguish it as a Google Cloud Storage connection rather than another type of connection. In addition, at least one of adminRoles, adminGroups, or adminUsers must be provided.

Java

Create a GCS connection

String adminRoleGuid = client.getRoleCache().getIdForName("$admin"); // (1)
Connection connection = Connection.creator( // (2)
        "gcs-connection", // (3)
        AtlanConnectorType.GCS, // (4)
        List.of(adminRoleGuid), // (5)
        List.of("group2"), // (6)
        List.of("jsmith")) // (7)
    .build();
AssetMutationResponse response = connection.save(client); // (8)
String connectionQualifiedName = response.getCreatedAssets().get(0).getQualifiedName(); // (9)

1. Retrieve the GUID for the admin role, to use later for defining the roles that can administer the connection.
2. Build up the minimum request to create a connection.
3. Provide a human-readable name for your connection, such as production or development.
4. Set the type of connection to GCS.
5. List the workspace roles that should be able to administer the connection (or null if none). All users with that workspace role (current and future) will be administrators of the connection. Note that the values here need to be the GUIDs of the workspace roles. At least one of adminRoles, adminGroups, or adminUsers must be provided.
6. List the group names that can administer this connection (or null if none). All users within that group (current and future) will be administrators of the connection. Note that the values here are the names of the groups. At least one of adminRoles, adminGroups, or adminUsers must be provided.
7. List the user names that can administer this connection (or null if none). Note that the values here are the usernames of the users. At least one of adminRoles, adminGroups, or adminUsers must be provided.
8. Actually call Atlan to create the connection. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.
9. Retrieve the qualifiedName for use in subsequent creation calls. (You'd probably want to do some null checking first.)

Python

Create a GCS connection

from pyatlan.client.atlan import AtlanClient
from pyatlan.model.assets import Connection, GCSBucket, GCSObject
from pyatlan.model.enums import AtlanConnectorType

client = AtlanClient()
admin_role_guid = client.role_cache.get_id_for_name("$admin") # (1)
connection = Connection.creator( # (2)
    client=client, # (3)
    name="gcs-connection", # (4)
    connector_type=AtlanConnectorType.GCS, # (5)
    admin_roles=[admin_role_guid], # (6)
    admin_groups=["group2"], # (7)
    admin_users=["jsmith"] # (8)
)

response = client.asset.save(connection) # (9)
connection_qualified_name = response.assets_created(asset_type=Connection)[0].qualified_name # (10)

1. Retrieve the GUID for the admin role, to use later for defining the roles that can administer the connection.
2. Build up the minimum request to create a connection.
3. You must provide a client instance.
4. Provide a human-readable name for your connection, such as production or development.
5. Set the type of connection to GCS.
6. List the workspace roles that should be able to administer the connection (or None if none). All users with that workspace role (current and future) will be administrators of the connection. Note that the values here need to be the GUIDs of the workspace roles. At least one of admin_roles, admin_groups, or admin_users must be provided.
7. List the group names that can administer this connection (or None if none). All users within that group (current and future) will be administrators of the connection. Note that the values here are the names of the groups. At least one of admin_roles, admin_groups, or admin_users must be provided.
8. List the user names that can administer this connection (or None if none). Note that the values here are the usernames of the users. At least one of admin_roles, admin_groups, or admin_users must be provided.
9. Actually call Atlan to create the connection.
10. Retrieve the qualifiedName for use in subsequent creation calls. (You'd probably want to do some null checking first.)

Kotlin

Create a GCS connection

val adminRoleGuid = client.roleCache.getIdForName("\$admin"); // (1)
val connection = Connection.creator( // (2)
        "gcs-connection", // (3)
        AtlanConnectorType.GCS, // (4)
        listOf(adminRoleGuid), // (5)
        listOf("group2"), // (6)
        listOf("jsmith")) // (7)
    .build()
val response = connection.save(client) // (8)
val connectionQualifiedName = response.createdAssets[0].qualifiedName // (9)

1. Retrieve the GUID for the admin role, to use later for defining the roles that can administer the connection.
2. Build up the minimum request to create a connection.
3. Provide a human-readable name for your connection, such as production or development.
4. Set the type of connection to GCS.
5. List the workspace roles that should be able to administer the connection (or null if none). All users with that workspace role (current and future) will be administrators of the connection. Note that the values here need to be the GUIDs of the workspace roles. At least one of adminRoles, adminGroups, or adminUsers must be provided.
6. List the group names that can administer this connection (or null if none). All users within that group (current and future) will be administrators of the connection. Note that the values here are the names of the groups. At least one of adminRoles, adminGroups, or adminUsers must be provided.
7. List the user names that can administer this connection (or null if none). Note that the values here are the usernames of the users. At least one of adminRoles, adminGroups, or adminUsers must be provided.
8. Actually call Atlan to create the connection. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.
9. Retrieve the qualifiedName for use in subsequent creation calls. (You'd probably want to do some null checking first.)

Raw REST API

POST /api/meta/entity/bulk

{
  "entities": [
    {
      "typeName": "Connection", // (1)
      "attributes": {
        "name": "gcs-connection", // (2)
        "connectorName": "gcs", // (3)
        "qualifiedName": "default/gcs/123456789", // (4)
        "category": "ObjectStore", // (5)
        "adminRoles": [ // (6)
          "e7ae0295-c60a-469a-bd2c-fb903943aa02"
        ],
        "adminGroups": [ // (7)
          "group2"
        ],
        "adminUsers": [ // (8)
          "jsmith"
        ]
      }
    }
  ]
}

1. The typeName must be exactly Connection.
2. Human-readable name for your connection, such as production or development.
3. The connectorName must be exactly gcs.
4. The qualifiedName should follow the pattern: default/gcs/<epoch>, where <epoch> is the time in milliseconds at which the connection is being created.
5. The category must be ObjectStore.
6. List any workspace roles that can administer this connection. All users with that workspace role (current and future) will be administrators of the connection. Note that the values here need to be the GUIDs of the workspace roles. At least one of adminRoles, adminGroups, or adminUsers must be provided.
7. List any groups that can administer this connection. All users within that group (current and future) will be administrators of the connection. Note that the values here are the names of the groups. At least one of adminRoles, adminGroups, or adminUsers must be provided.
8. List any users that can administer this connection. Note that the values here are the usernames of the users. At least one of adminRoles, adminGroups, or adminUsers must be provided.

Access policies

Atlan creates the policies that grant access to a connection, including the ability to retrieve the connection and to create assets within it, asynchronously. It can take several seconds (even up to approximately 30 seconds) before these are in place after creating the connection.

You may therefore need to wait before you'll be able to create the assets below within the connection.

To confirm access, retrieve the connection after it has been created. The SDKs' retry loops will automatically retry until the connection can be successfully retrieved. At that point, your API token has permission to create the other assets.

Note: if you are reusing an existing connection rather than creating one via your API token, you must give your API token a persona that has access to that connection. Otherwise all attempts to create, read, update, or delete assets within that connection will fail due to a lack of permissions.

GCSBucket

A Google Cloud Storage bucket requires a name and a qualifiedName. For creation, you also need to specify the connectionQualifiedName of the connection for the bucket.

Java

Create a GCS bucket

GCSBucket gcsBucket = GCSBucket.creator( // (1)
        "mybucket", // (2)
        connectionQualifiedName) // (3)
    .gcsObjectCount(10) // (4)
    .build();
AssetMutationResponse response = gcsBucket.save(client); // (5)
gcsBucket = response.getResult(gcsBucket); // (6)

1. Build up the minimum request to create a bucket.
2. Provide a human-readable name for your bucket.
3. Provide the qualifiedName of the connection for this bucket.
4. (Optional) To make sure the UI displays the correct count of GCSObject's, set the gcsObjectCount directly on the GCSBucket instance.
5. Actually call Atlan to create the bucket. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.
6. Retrieve the created bucket for use in subsequent creation calls. (You'd probably want to do some null checking first.)

Python

Create a GCS bucket

gcsBucket = GCSBucket.creator( # (1)
    name = "mybucket", # (2)
    connection_qualified_name = connection_qualified_name # (3)
)
gcsBucket.gcs_object_count = 10 # (4)
response = client.asset.save(gcsBucket) # (5)
gcs_bucket_qualifed_name = response.assets_created(asset_type=GCSBucket)[0].qualified_name # (6)

1. Build up the minimum request to create a bucket.
2. Provide a human-readable name for your bucket.
3. Provide the qualified_name of the connection for this bucket.
4. (Optional) To make sure the UI displays the correct count of GCSObject's, set the gcs_object_count directly on the GCSBucket instance.
5. Actually call Atlan to create the bucket.
6. Retrieve the created bucket for use in subsequent creation calls.(You'd probably want to do some null checking first.)

Kotlin

Create a GCS bucket

var gcsBucket = GCSBucket.creator( // (1)
        "mybucket", // (2)
        connectionQualifiedName) // (3)
    .gcsObjectCount(10) // (4)
    .build()
val response = gcsBucket.save(client) // (5)
gcsBucket = response.getResult(gcsBucket) // (6)

1. Build up the minimum request to create a bucket.
2. Provide a human-readable name for your bucket.
3. Provide the qualifiedName of the connection for this bucket.
4. (Optional) To make sure the UI displays the correct count of GCSObject's, set the gcsObjectCount directly on the GCSBucket instance.
5. Actually call Atlan to create the bucket. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.
6. Retrieve the created bucket for use in subsequent creation calls. (You'd probably want to do some null checking first.)

Raw REST API

POST /api/meta/entity/bulk

{
  "entities": [
    {
      "typeName": "GCSBucket", // (1)
      "attributes": {
        "name": "mybucket", // (2)
        "qualifiedName": "default/gcs/123456789/mybucket", // (3)
        "connectionQualifiedName": "default/gcs/123456789", // (4)
        "connectorName": "gcs" // (5)
      }
    }
  ]
}

1. The typeName must be exactly GCSBucket.
2. Human-readable name for your bucket.
3. The qualifiedName should follow the pattern: default/gcs/<epoch>/<name>, where default/gcs/<epoch> is the qualifiedName of the connection for this bucket and <name> is the unique name for this bucket.
4. The connectionQualifiedName must be the exact qualifiedName of the connection for this bucket.
5. The connectorName must be exactly gcs.

GCSObject

A Google Cloud Storage object requires a name and a qualifiedName. For creation, you also need to specify the bucketQualifiedName of the bucket that will contain the object.

Java

Create a GCS object

GCSObject gcsObject = GCSObject.creator( // (1)
        "myobject.csv", // (2)
        gcsBucket) // (3)
    .build();
AssetMutationResponse response = gcsObject.save(client); // (4)

1. Build up the minimum request to create an object.
2. Provide a human-readable name for your object.
3. Provide the bucket for this object. If you didn't already have the bucket, you could also use GCSBucket.refByGuid() with the GUID of the bucket, or GCSBucket.refByQualifiedName() with the qualifiedName of the bucket.
4. Actually call Atlan to create the object. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.

Python

Create a GCS object

gcsObject = GCSObject.creator( # (1)
    name="myobject.csv", # (2)
    gcs_bucket_name=bucket_name, # (3)
    gcs_bucket_qualified_name=gcs_bucket_qualified_name # (4)
)
response = client.asset.save(gcsObject) # (5)

1. Build up the minimum request to create an object.
2. Provide a human-readable name for your object.
3. Provide the name of the bucket this object should be created within.
4. Provide the qualified_name of the bucket for this object.
5. Actually call Atlan to create the object.

Kotlin

Create a GCS object

val gcsObject = GCSObject.creator( // (1)
        "myobject.csv", // (2)
        gcsBucket) // (3)
    .build()
val response = gcsObject.save(client) // (4)

1. Build up the minimum request to create an object.
2. Provide a human-readable name for your object.
3. Provide the bucket for this object. If you didn't already have the bucket, you could also use GCSBucket.refByGuid() with the GUID of the bucket, or GCSBucket.refByQualifiedName() with the qualifiedName of the bucket.
4. Actually call Atlan to create the object. Because this operation will persist the asset in Atlan, you must provide it an AtlanClient through which to connect to the tenant.

Raw REST API

POST /api/meta/entity/bulk

{
  "entities": [
    {
      "typeName": "GCSObject", // (1)
      "attributes": {
        "name": "myobject.csv", // (2)
        "qualifiedName": "default/gcs/123456789/mybucket/myobject.csv", // (3)
        "connectionQualifiedName": "default/gcs/123456789", // (4)
        "connectorName": "gcs", // (5)
        "gcsBucket": { // (6)
          "typeName": "GSCBucket", // (7)
          "uniqueAttributes": { // (8)
            "qualifiedName": "default/gcs/123456789/mybucket"
          }
        },
        "gcsBucketName": "mybucket", // (9)
        "gcsBucketQualifiedName": "default/gcs/123456789/mybucket" // (10)
      }
    }
  ]
}

1. The typeName must be exactly GCSObject.
2. Human-readable name for your object.
3. The qualifiedName should follow the pattern: default/gcs/<epoch>/<bucket>/<name>, where default/gcs/<epoch>/<bucket> is the qualifiedName of the bucket that contains this object and <name> is the unique name for this object.
4. The connectionQualifiedName must be the exact qualifiedName of the connection for this object.
5. The connectorName must be exactly gcs.
6. The bucket in which this object exists is embedded in the gcsBucket attribute.
7. The typeName for this embedded reference must be GCSBucket.
8. To complete the reference, you must include a uniqueAttributes object with the qualifiedName of the bucket. Note: the bucket must already exist in Atlan before creating the object.
9. The gcsBucketName should be the human-readable name of the bucket.
10. The gcsBucketQualifiedName should be the qualifiedName of the bucket.

By prefix

Create a GCS object using prefix:

Java

Coming soon

Python

Create a GCS object using prefix

gcs_object = GCSObject.creator_with_prefix(  # (1)
    name="myobject.csv",  # (2)
    connection_qualified_name=connection_qualified_name,  # (3)
    prefix="/some/folder/structure",  # (4)
    gcs_bucket_name=bucket_name, # (5)
    gcs_bucket_qualified_name=bucket_qualified_name,  # (6)
)
response = client.asset.save(gcs_object)  # (7)

1. Build up the minimum request to create an object.
2. Provide a human-readable name for your object.
3. Provide the qualified_name of the connection for this object.
4. Provide the folder path where the object is located within the bucket.
5. Provide the name of the bucket this object should be created within.
6. Provide the qualified_name of the bucket this object should be created within.
7. Actually call Atlan to create the object.

Java

Coming soon

Raw REST API

POST /api/meta/entity/bulk

{
  "entities": [
    {
      "typeName": "GCSObject", // (1)
      "attributes": {
        "name": "myobject.csv", // (2)
        "qualifiedName": "default/gcs/123456789/mybucket//some/folder/structure/myobject.csv", // (3)
        "gcsObjectKey": "/some/folder/structure/myobject.csv", // (4)
        "connectionQualifiedName": "default/gcs/123456789", // (5)
        "connectorName": "gcs", // (6)
        "gcsBucket": { // (7)
          "typeName": "GSCBucket", // (8)
          "uniqueAttributes": { // (9)
            "qualifiedName": "default/gcs/123456789/mybucket"
          }
        },
        "gcsBucketName": "mybucket", // (10)
        "gcsBucketQualifiedName": "default/gcs/123456789/mybucket" // (11)
      }
    }
  ]
}

1. The typeName must be exactly GCSObject.
2. Human-readable name for your object.
3. The qualifiedName should follow the pattern: default/gcs/<epoch>/&lt;bucket-name&gt;/<prefix>/<name>, where default/gcs/<epoch> is the qualifiedName of the connection for this object, and <prefix>/<name> is the folder path where this object is located within the bucket.
4. Provide the folder path where the object is located within the bucket.
5. The connectionQualifiedName must be the exact qualifiedName of the connection for this object.
6. The connectorName must be exactly gcs.
7. The bucket in which this object exists is embedded in the gcsBucket attribute.
8. The typeName for this embedded reference must be GCSBucket.
9. To complete the reference, you must include a uniqueAttributes object with the qualifiedName of the bucket. Note: the bucket must already exist in Atlan before creating the object.
10. The gcsBucketName should be the human-readable name of the bucket.
11. The gcsBucketQualifiedName should be the qualifiedName of the bucket.

Available relationships

Every level of the object store structure is an Asset, and can therefore be related to the following other assets.

Footnotes

1. Although if you want to delete everything in a connection, your better avenue is the packaged connection delete utility in the UI. ↩