feat: resource tags in dataset

google/cloud/bigquery/dataset.py

@@ -530,6 +530,7 @@ class Dataset(object):
         "storage_billing_model": "storageBillingModel",
         "max_time_travel_hours": "maxTimeTravelHours",
         "default_rounding_mode": "defaultRoundingMode",
+        "resource_tags": "resourceTags",
     }
 
     def __init__(self, dataset_ref) -> None:
@@ -801,6 +802,28 @@ def labels(self, value):
             raise ValueError("Pass a dict")
         self._properties["labels"] = value
 
+    @property
+    def resource_tags(self):
+        """Dict[str, str]: Resource tags of the dataset.
+
+        Optional. The tags attached to this dataset. Tag keys are globally
+        unique. Tag key is expected to be in the namespaced format, for
+        example "123456789012/environment" where 123456789012 is
+        the ID of the parent organization or project resource for this tag
+        key. Tag value is expected to be the short name, for example
+        "Production".
+
+        Raises:
+            ValueError: for invalid value types.
+        """
+        return self._properties.setdefault("resourceTags", {})
+
+    @resource_tags.setter
+    def resource_tags(self, value):
+        if not isinstance(value, dict) and value is not None:
+            raise ValueError("Pass a dict")
+        self._properties["resourceTags"] = value
+
     @property
     def default_encryption_configuration(self):
         """google.cloud.bigquery.encryption_configuration.EncryptionConfiguration: Custom

noxfile.py

@@ -219,6 +219,9 @@ def system(session):
     # Data Catalog needed for the column ACL test with a real Policy Tag.
     session.install("google-cloud-datacatalog", "-c", constraints_path)
 
+    # Resource Manager needed for test with a real Resource Tag.
+    session.install("google-cloud-resource-manager", "-c", constraints_path)
+
     if session.python in ["3.11", "3.12"]:
         extras = "[bqstorage,ipywidgets,pandas,tqdm,opentelemetry]"
     else:
@@ -366,6 +369,7 @@ def prerelease_deps(session):
     session.install(
         "freezegun",
         "google-cloud-datacatalog",
+        "google-cloud-resource-manager",
         "google-cloud-storage",
         "google-cloud-testutils",
         "psutil",

tests/system/test_client.py

@@ -45,6 +45,8 @@
 from google.cloud import storage
 from google.cloud.datacatalog_v1 import types as datacatalog_types
 from google.cloud.datacatalog_v1 import PolicyTagManagerClient
+from google.cloud.resourcemanager_v3 import types as resourcemanager_types
+from google.cloud.resourcemanager_v3 import TagKeysClient, TagValuesClient
 import psutil
 import pytest
 from test_utils.retry import RetryErrors
@@ -159,6 +161,8 @@ def setUp(self):
 
     def tearDown(self):
         policy_tag_client = PolicyTagManagerClient()
+        tag_keys_client = TagKeysClient()
+        tag_values_client = TagValuesClient()
 
         def _still_in_use(bad_request):
             return any(
@@ -178,6 +182,10 @@ def _still_in_use(bad_request):
                 retry_in_use(Config.CLIENT.delete_table)(doomed)
             elif isinstance(doomed, datacatalog_types.Taxonomy):
                 policy_tag_client.delete_taxonomy(name=doomed.name)
+            elif isinstance(doomed, resourcemanager_types.TagKey):
+                tag_keys_client.delete_tag_key(name=doomed.name).result()
+            elif isinstance(doomed, resourcemanager_types.TagValue):
+                tag_values_client.delete_tag_value(name=doomed.name).result()
             else:
                 doomed.delete()
 
@@ -284,27 +292,41 @@ def test_update_dataset(self):
         self.assertIsNone(dataset.friendly_name)
         self.assertIsNone(dataset.description)
         self.assertEqual(dataset.labels, {})
+        self.assertEqual(dataset.resource_tags, {})
         self.assertIs(dataset.is_case_insensitive, False)
 
         dataset.friendly_name = "Friendly"
         dataset.description = "Description"
         dataset.labels = {"priority": "high", "color": "blue"}
+        dataset.resource_tags =  {"123456789012/env": "prod", "123456789012/component": "batch"}
         dataset.is_case_insensitive = True
         ds2 = Config.CLIENT.update_dataset(
-            dataset, ("friendly_name", "description", "labels", "is_case_insensitive")
+            dataset, ("friendly_name", "description", "labels", "resource_tags", "is_case_insensitive")
         )
         self.assertEqual(ds2.friendly_name, "Friendly")
         self.assertEqual(ds2.description, "Description")
         self.assertEqual(ds2.labels, {"priority": "high", "color": "blue"})
+        self.assertEqual(ds2.resource_tags, {"123456789012/env": "prod", "123456789012/component": "batch"})
         self.assertIs(ds2.is_case_insensitive, True)
 
         ds2.labels = {
             "color": "green",  # change
             "shape": "circle",  # add
             "priority": None,  # delete
         }
-        ds3 = Config.CLIENT.update_dataset(ds2, ["labels"])
+        ds2.resource_tags = {
+            "123456789012/env": "dev",          # change
+            "123456789012/project": "atlas",    # add
+            "123456789012/component": None,     # delete
+        }
+        ds3 = Config.CLIENT.update_dataset(ds2, ["labels", "resource_tags"])
         self.assertEqual(ds3.labels, {"color": "green", "shape": "circle"})
+        self.assertEqual(ds3.resource_tags, {"123456789012/env": "dev", "123456789012/project": "atlas"})
+
+        # Remove all tags
+        ds3.resource_tags = None
+        ds4 = Config.CLIENT.update_table(ds3, ["resource_tags"])
+        self.assertEqual(ds4.resource_tags, {})
 
         # If we try to update using d2 again, it will fail because the
         # previous update changed the ETag.

tests/unit/test_client.py

@@ -2028,6 +2028,7 @@ def test_update_dataset(self):
         LABELS = {"priority": "high"}
         ACCESS = [{"role": "OWNER", "userByEmail": "[email protected]"}]
         EXP = 17
+        RESOURCE_TAGS = {"123456789012/key": "value"}
         RESOURCE = {
             "datasetReference": {"projectId": self.PROJECT, "datasetId": self.DS_ID},
             "etag": "etag",
@@ -2037,6 +2038,7 @@ def test_update_dataset(self):
             "defaultTableExpirationMs": EXP,
             "labels": LABELS,
             "access": ACCESS,
+            "resourceTags": RESOURCE_TAGS,
         }
         creds = _make_credentials()
         client = self._make_one(project=self.PROJECT, credentials=creds)
@@ -2048,12 +2050,14 @@ def test_update_dataset(self):
         ds.default_table_expiration_ms = EXP
         ds.labels = LABELS
         ds.access_entries = [AccessEntry("OWNER", "userByEmail", "[email protected]")]
+        ds.resource_tags = RESOURCE_TAGS
         fields = [
             "description",
             "friendly_name",
             "location",
             "labels",
             "access_entries",
+            "resource_tags",
         ]
 
         with mock.patch(
@@ -2077,6 +2081,7 @@ def test_update_dataset(self):
                 "location": LOCATION,
                 "labels": LABELS,
                 "access": ACCESS,
+                "resourceTags": RESOURCE_TAGS,
             },
             path="/" + PATH,
             timeout=7.5,
@@ -2086,6 +2091,7 @@ def test_update_dataset(self):
         self.assertEqual(ds2.location, ds.location)
         self.assertEqual(ds2.labels, ds.labels)
         self.assertEqual(ds2.access_entries, ds.access_entries)
+        self.assertEqual(ds2.resource_tags, ds.resource_tags)
 
         # ETag becomes If-Match header.
         ds._properties["etag"] = "etag"

tests/unit/test_create_dataset.py

@@ -65,6 +65,7 @@ def test_create_dataset_w_attrs(client, PROJECT, DS_ID):
         "tableId": "northern-hemisphere",
     }
     DEFAULT_ROUNDING_MODE = "ROUND_HALF_EVEN"
+    RESOURCE_TAGS = {"123456789012/foo": "bar"}
     RESOURCE = {
         "datasetReference": {"projectId": PROJECT, "datasetId": DS_ID},
         "etag": "etag",
@@ -76,6 +77,7 @@ def test_create_dataset_w_attrs(client, PROJECT, DS_ID):
         "labels": LABELS,
         "access": [{"role": "OWNER", "userByEmail": USER_EMAIL}, {"view": VIEW}],
         "defaultRoundingMode": DEFAULT_ROUNDING_MODE,
+        "resourceTags": RESOURCE_TAGS,
     }
     conn = client._connection = make_connection(RESOURCE)
     entries = [
@@ -91,6 +93,7 @@ def test_create_dataset_w_attrs(client, PROJECT, DS_ID):
     before.default_table_expiration_ms = 3600
     before.location = LOCATION
     before.labels = LABELS
+    before.resource_tags = RESOURCE_TAGS
     before.default_rounding_mode = DEFAULT_ROUNDING_MODE
     after = client.create_dataset(before)
     assert after.dataset_id == DS_ID
@@ -103,6 +106,7 @@ def test_create_dataset_w_attrs(client, PROJECT, DS_ID):
     assert after.default_table_expiration_ms == 3600
     assert after.labels == LABELS
     assert after.default_rounding_mode == DEFAULT_ROUNDING_MODE
+    assert after.resource_tags == RESOURCE_TAGS
 
     conn.api_request.assert_called_once_with(
         method="POST",
@@ -119,6 +123,7 @@ def test_create_dataset_w_attrs(client, PROJECT, DS_ID):
                 {"view": VIEW, "role": None},
             ],
             "labels": LABELS,
+            "resourceTags": RESOURCE_TAGS,
         },
         timeout=DEFAULT_TIMEOUT,
     )

tests/unit/test_dataset.py

@@ -894,6 +894,28 @@ def test_location_setter(self):
         dataset.location = "LOCATION"
         self.assertEqual(dataset.location, "LOCATION")
 
+    def test_resource_tags_update_in_place(self):
+        dataset = self._make_one(self.DS_REF)
+        tags = dataset.resource_tags
+        tags["123456789012/foo"] = "bar"  # update in place
+        self.assertEqual(dataset.resource_tags, {"123456789012/foo": "bar"})
+
+    def test_resource_tags_setter(self):
+        dataset = self._make_one(self.DS_REF)
+        dataset.resource_tags = {"123456789012/foo": "bar"}
+        self.assertEqual(dataset.resource_tags, {"123456789012/foo": "bar"})
+
+    def test_resource_tags_setter_bad_value(self):
+        dataset = self._make_one(self.DS_REF)
+        with self.assertRaises(ValueError):
+            dataset.resource_tags = "invalid"
+        with self.assertRaises(ValueError):
+            dataset.resource_tags = 123
+
+    def test_resource_tags_getter_missing_value(self):
+        dataset = self._make_one(self.DS_REF)
+        self.assertEqual(dataset.resource_tags, {})
+
     def test_labels_update_in_place(self):
         dataset = self._make_one(self.DS_REF)
         del dataset._properties["labels"]  # don't start w/ existing dict