feat: add support for proto3 optional tag (#727)

* feat: add support for proto3 optional tag

* format writer.py

* Add the same changes to v1beta2

* Add systen test for proto3 support

* Remove v1beta2 modifications

* Fix issue in the test script and reformat

* fix lint

* Fix typo in the test

* Remove unneed offset

---------

Co-authored-by: Lingqing Gan <lingqing.gan@gmail.com>

Author: 2 people

packages/google-cloud-bigquery-storage/google/cloud/bigquery_storage_v1/writer.py

@@ -100,6 +100,17 @@ def __init__(
         # The threads created in ``._open()``.
         self._consumer = None
 
+        # The protobuf payload will be decoded as proto2 on the server side. The schema is also
+        # specified as proto2. Hence we must clear proto3-only features. This works since proto2 and
+        # proto3 are binary-compatible.
+        proto_descriptor = (
+            self._inital_request_template.proto_rows.writer_schema.proto_descriptor
+        )
+        for field in proto_descriptor.field:
+            field.ClearField("oneof_index")
+            field.ClearField("proto3_optional")
+        proto_descriptor.ClearField("oneof_decl")
+
     @property
     def is_active(self) -> bool:
         """bool: True if this manager is actively streaming.

packages/google-cloud-bigquery-storage/tests/system/test_writer.py

@@ -16,6 +16,29 @@
 import pytest
 
 from google.cloud.bigquery_storage_v1 import types as gapic_types
+from google.cloud.bigquery_storage_v1.writer import AppendRowsStream
+import uuid
+
+
+@pytest.fixture
+def table(project_id, dataset, bq_client):
+    from google.cloud import bigquery
+
+    schema = [
+        bigquery.SchemaField("first_name", "STRING", mode="NULLABLE"),
+        bigquery.SchemaField("last_name", "STRING", mode="NULLABLE"),
+        bigquery.SchemaField("age", "INTEGER", mode="NULLABLE"),
+    ]
+
+    unique_suffix = str(uuid.uuid4()).replace("-", "_")
+    table_id = "users_" + unique_suffix
+    table_id_full = f"{project_id}.{dataset.dataset_id}.{table_id}"
+    bq_table = bigquery.Table(table_id_full, schema=schema)
+    created_table = bq_client.create_table(bq_table)
+
+    yield created_table
+
+    bq_client.delete_table(created_table)
 
 
 @pytest.fixture(scope="session")
@@ -31,3 +54,63 @@ def test_append_rows_with_invalid_stream_name_fails_fast(bqstorage_write_client)
 
     with pytest.raises(exceptions.GoogleAPICallError):
         bqstorage_write_client.append_rows(bad_request)
+
+
+def test_append_rows_with_proto3(bqstorage_write_client, table):
+    import proto
+    from google.protobuf import descriptor_pb2
+
+    # Using Proto Plus to build proto3
+    # Declare proto3 field `optional` for presence
+    class PersonProto(proto.Message):
+        first_name = proto.Field(
+            proto.STRING,
+            number=1,
+            optional=True,
+        )
+        last_name = proto.Field(
+            proto.STRING,
+            number=2,
+            optional=True,
+        )
+        age = proto.Field(
+            proto.INT64,
+            number=3,
+            optional=True,
+        )
+
+    person_pb = PersonProto.pb()
+
+    stream_name = f"projects/{table.project}/datasets/{table.dataset_id}/tables/{table.table_id}/_default"
+    request_template = gapic_types.AppendRowsRequest()
+    request_template.write_stream = stream_name
+
+    proto_schema = gapic_types.ProtoSchema()
+    proto_descriptor = descriptor_pb2.DescriptorProto()
+    person_pb.DESCRIPTOR.CopyToProto(
+        proto_descriptor,
+    )
+    proto_schema.proto_descriptor = proto_descriptor
+    proto_data = gapic_types.AppendRowsRequest.ProtoData()
+    proto_data.writer_schema = proto_schema
+    request_template.proto_rows = proto_data
+
+    append_rows_stream = AppendRowsStream(
+        bqstorage_write_client,
+        request_template,
+    )
+
+    request = gapic_types.AppendRowsRequest()
+    proto_data = gapic_types.AppendRowsRequest.ProtoData()
+    proto_rows = gapic_types.ProtoRows()
+    row = person_pb()
+    row.first_name = "fn"
+    row.last_name = "ln"
+    row.age = 20
+    proto_rows.serialized_rows.append(row.SerializeToString())
+    proto_data.rows = proto_rows
+    request.proto_rows = proto_data
+    response_future = append_rows_stream.send(request)
+
+    assert response_future.result()
+    # The request should success