[source-google-analytics-data-api] - Migrate to CDK v5 (#47013)

airbyte-integrations/connectors/source-google-analytics-data-api/metadata.yaml

@@ -12,7 +12,7 @@ data:
   connectorSubtype: api
   connectorType: source
   definitionId: 3cc2eafd-84aa-4dca-93af-322d9dfeec1a
-  dockerImageTag: 2.5.13
+  dockerImageTag: 2.6.0
   dockerRepository: airbyte/source-google-analytics-data-api
   documentationUrl: https://docs.airbyte.com/integrations/sources/google-analytics-data-api
   githubIssueLabel: source-google-analytics-data-api

airbyte-integrations/connectors/source-google-analytics-data-api/pyproject.toml

@@ -3,7 +3,7 @@ requires = [ "poetry-core>=1.0.0",]
 build-backend = "poetry.core.masonry.api"
 
 [tool.poetry]
-version = "2.5.13"
+version = "2.6.0"
 name = "source-google-analytics-data-api"
 description = "Source implementation for Google Analytics Data Api."
 authors = [ "Airbyte <[email protected]>",]
@@ -16,12 +16,12 @@ repository = "https://github.com/airbytehq/airbyte"
 include = "source_google_analytics_data_api"
 
 [tool.poetry.dependencies]
-python = "^3.9,<3.12"
+python = "^3.10,<3.12"
 cryptography = "==42.0.5"
 requests = "==2.31.0"
-airbyte-cdk = "^3"
+airbyte-cdk = "^5.0.0"
 PyJWT = "==2.8.0"
-pandas = "==2.2.0"
+pandas = "^2.2.0"
 
 [tool.poetry.scripts]
 source-google-analytics-data-api = "source_google_analytics_data_api.run:run"

airbyte-integrations/connectors/source-google-analytics-data-api/source_google_analytics_data_api/config_migrations.py

@@ -7,8 +7,10 @@
 from typing import Any, List, Mapping
 
 import dpath.util
+import orjson
 from airbyte_cdk.config_observation import create_connector_config_control_message
 from airbyte_cdk.entrypoint import AirbyteEntrypoint
+from airbyte_cdk.models import AirbyteMessageSerializer
 from airbyte_cdk.sources.message import InMemoryMessageRepository, MessageRepository
 
 from .source import SourceGoogleAnalyticsDataApi
@@ -67,7 +69,7 @@ def _emit_control_message(cls, migrated_config: Mapping[str, Any]) -> None:
         cls.message_repository.emit_message(create_connector_config_control_message(migrated_config))
         # emit the Airbyte Control Message from message queue to stdout
         for message in cls.message_repository.consume_queue():
-            print(message.json(exclude_unset=True))
+            print(orjson.dumps(AirbyteMessageSerializer.dump(message)).decode())
 
     @classmethod
     def migrate(cls, args: List[str], source: SourceGoogleAnalyticsDataApi) -> None:
@@ -145,7 +147,7 @@ def _emit_control_message(cls, migrated_config: Mapping[str, Any]) -> None:
         cls.message_repository.emit_message(create_connector_config_control_message(migrated_config))
         # emit the Airbyte Control Message from message queue to stdout
         for message in cls.message_repository.consume_queue():
-            print(message.json(exclude_unset=True))
+            print(orjson.dumps(AirbyteMessageSerializer.dump(message)).decode())
 
     @classmethod
     def migrate(cls, args: List[str], source: SourceGoogleAnalyticsDataApi) -> None:
@@ -215,7 +217,7 @@ def _emit_control_message(cls, migrated_config: Mapping[str, Any]) -> None:
         cls.message_repository.emit_message(create_connector_config_control_message(migrated_config))
         # emit the Airbyte Control Message from message queue to stdout
         for message in cls.message_repository.consume_queue():
-            print(message.json(exclude_unset=True))
+            print(orjson.dumps(AirbyteMessageSerializer.dump(message)).decode())
 
     @classmethod
     def migrate(cls, args: List[str], source: SourceGoogleAnalyticsDataApi) -> None:

airbyte-integrations/connectors/source-google-analytics-data-api/source_google_analytics_data_api/source.py

@@ -2,6 +2,7 @@
 # Copyright (c) 2023 Airbyte, Inc., all rights reserved.
 #
 
+import copy
 import datetime
 import json
 import logging
@@ -354,10 +355,14 @@ def request_body_json(
         stream_slice: Mapping[str, Any] = None,
         next_page_token: Mapping[str, Any] = None,
     ) -> Optional[Mapping]:
+        if stream_slice and "startDate" in stream_slice and "endDate" in stream_slice:
+            date_range = {"startDate": stream_slice["startDate"], "endDate": stream_slice["endDate"]}
+        else:
+            date_range = stream_slice
         payload = {
             "metrics": [{"name": m} for m in self.config["metrics"]],
             "dimensions": [{"name": d} for d in self.config["dimensions"]],
-            "dateRanges": [stream_slice],
+            "dateRanges": [date_range],
             "returnPropertyQuota": True,
             "offset": str(0),
             "limit": str(self.page_size),
@@ -414,7 +419,6 @@ def request_body_json(
         next_page_token: Mapping[str, Any] = None,
     ) -> Optional[Mapping]:
         payload = super().request_body_json(stream_state, stream_slice, next_page_token)
-
         # remove offset and limit fields according to their absence in
         # https://developers.google.com/analytics/devguides/reporting/data/v1/rest/v1beta/properties/runPivotReport
         payload.pop("offset", None)

airbyte-integrations/connectors/source-google-analytics-data-api/unit_tests/test_source.py

@@ -5,10 +5,9 @@
 from unittest.mock import MagicMock, patch
 
 import pytest
-from airbyte_cdk.models import AirbyteConnectionStatus, FailureType
+from airbyte_cdk.models import AirbyteConnectionStatus, FailureType, Status
 from airbyte_cdk.sources.streams.http.http import HttpStatusErrorHandler
 from airbyte_cdk.utils import AirbyteTracedException
-from airbyte_protocol.models import Status
 from source_google_analytics_data_api import SourceGoogleAnalyticsDataApi
 from source_google_analytics_data_api.api_quota import GoogleAnalyticsApiQuotaBase
 from source_google_analytics_data_api.source import GoogleAnalyticsDatApiErrorHandler, MetadataDescriptor

airbyte-integrations/connectors/source-google-analytics-data-api/unit_tests/test_streams.py

@@ -64,7 +64,8 @@ def test_request_params(patch_base_class):
 
 
 def test_request_body_json(patch_base_class):
-    request_body_params = {"stream_state": MagicMock(), "stream_slice": MagicMock(), "next_page_token": None}
+    stream_slice = {"startDate": "2024-01-01", "endDate": "2024-01-31"}
+    request_body_params = {"stream_state": MagicMock(), "stream_slice": stream_slice, "next_page_token": None}
 
     expected_body_json = {
         "metrics": [
@@ -84,7 +85,10 @@ def test_request_body_json(patch_base_class):
             {"name": "browser"},
         ],
         "keepEmptyRows": True,
-        "dateRanges": [request_body_params["stream_slice"]],
+        "dateRanges": [{
+            "startDate": request_body_params["stream_slice"]["startDate"],
+            "endDate": request_body_params["stream_slice"]["endDate"]
+        }],
         "returnPropertyQuota": True,
         "offset": str(0),
         "limit": "100000",

docs/integrations/sources/google-analytics-data-api.md

@@ -272,6 +272,7 @@ The Google Analytics connector is subject to Google Analytics Data API quotas. P
 
 | Version | Date       | Pull Request                                             | Subject                                                                                |
 |:--------|:-----------| :------------------------------------------------------- |:---------------------------------------------------------------------------------------|
+| 2.6.0 | 2024-10-28 | [47013](https://github.com/airbytehq/airbyte/pull/47013) | Migrate to CDK v5 |
 | 2.5.13 | 2024-10-28 | [47061](https://github.com/airbytehq/airbyte/pull/47061) | Update dependencies |
 | 2.5.12 | 2024-10-12 | [46819](https://github.com/airbytehq/airbyte/pull/46819) | Update dependencies |
 | 2.5.11 | 2024-10-05 | [46475](https://github.com/airbytehq/airbyte/pull/46475) | Update dependencies |