Include user agent for pseudoanonymized usage measurement (#491)

Adds report_usage_id variable to jobs service.
Uses this variable to determine whether or not to include client_info when creating bigquery client.
Add unit test to check the opt-in/out works as expected for the client_info.

backend/jobs/workers/bigquery/bq_worker.py

@@ -1,4 +1,4 @@
-# Copyright 2020 Google Inc. All rights reserved.
+# Copyright 2024 Google Inc. All rights reserved.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -14,11 +14,11 @@
 
 """CRMint's abstract worker dealing with BigQuery."""
 
-
+import os
 import time
 
+from google.api_core.client_info import ClientInfo
 from google.cloud import bigquery
-
 from jobs.workers import worker
 
 
@@ -42,7 +42,16 @@ class BQWorker(worker.Worker):
   ]
 
   def _get_client(self):
-    return bigquery.Client(client_options={'scopes': self._SCOPES})
+    client_info = None
+    if 'REPORT_USAGE_ID' in os.environ:
+      client_id = os.getenv('REPORT_USAGE_ID')
+      opt_out = not bool(client_id)
+      if not opt_out:
+        client_info = ClientInfo(user_agent='cloud-solutions/crmint-usage-v3')
+    return bigquery.Client(
+      client_options={'scopes': self._SCOPES},
+      client_info=client_info,
+    )
 
   def _get_prefix(self):
     return f'{self._pipeline_id}_{self._job_id}_{self.__class__.__name__}'

backend/tests/jobs/unit/workers/bq_worker_tests.py

@@ -1,11 +1,14 @@
 """Tests for bq_worker."""
 
 from unittest import mock
+import os
 
 from absl.testing import absltest
 from absl.testing import parameterized
+
 from google.auth import credentials
 from google.cloud import bigquery
+from google.api_core.client_info import ClientInfo
 
 from jobs.workers import worker
 from jobs.workers.bigquery import bq_worker
@@ -82,6 +85,49 @@ def test_generates_proper_bq_table_name_from_params(self):
     self.assertEqual('a_project.a_dataset_id.a_table_id',
                      worker._generate_qualified_bq_table_name())
 
+class BQWorkerGetClientTest(parameterized.TestCase):
+
+  @parameterized.parameters(
+      {
+        'report_usage_id_present': True,
+        'client_info_user_agent': 'cloud-solutions/crmint-usage-v3',
+      },
+      {
+        'report_usage_id_present': False, 
+        'client_info_user_agent': None
+      },
+  )
+  def test_get_client_handles_report_usage_id(
+      self, report_usage_id_present, client_info_user_agent):
+    report_usage_id = 'some-usage-id' if report_usage_id_present else ''
+    with (
+        mock.patch.dict(
+            os.environ,
+            {'REPORT_USAGE_ID': report_usage_id}
+            if report_usage_id_present
+            else {},
+        ),
+        mock.patch('os.getenv', return_value=report_usage_id) as getenv_mock,
+        mock.patch('google.cloud.bigquery.Client') as client_mock,
+    ):
+      worker_inst = bq_worker.BQWorker({}, 0, 0)
+      worker_inst._get_client()
+
+      if report_usage_id_present:
+        getenv_mock.assert_called_with('REPORT_USAGE_ID')
+      else:
+        getenv_mock.assert_not_called()
+
+      client_mock.assert_called_once()
+      _, kwargs = client_mock.call_args
+      if report_usage_id_present:
+        self.assertIsInstance(kwargs['client_info'], ClientInfo)
+        self.assertEqual(
+          kwargs['client_info'].user_agent, client_info_user_agent
+        )
+      else:
+        self.assertIsNone(kwargs.get('client_info'))
+
 
 if __name__ == '__main__':
   absltest.main()

cli/appcli.py

@@ -54,7 +54,7 @@ def _ask_permission(self):
         fg='yellow')
     msg += click.style(pkg_name, fg='red', bold=True)
     msg += click.style(
-        ' better! \nMay we anonymously report usage statistics to improve the'
+        ' better! \nMay we anonymously report usage statistics to improve the '
         'tool over time? \nMore info: https://github.com/google/crmint & '
         'https://google.github.io/crmint',
         fg='yellow')

terraform/services.tf

@@ -209,6 +209,10 @@ resource "google_cloud_run_service" "jobs_run" {
           name  = "PUBSUB_VERIFICATION_TOKEN"
           value = random_id.pubsub_verification_token.b64_url
         }
+        env {
+          name  = "REPORT_USAGE_ID"
+          value = var.report_usage_id
+        }
       }
 
       timeout_seconds = 900  # 15min