resolve bug with daily notice metadata stores not days without notices in TED API

dags/daily_notices_metadata_update.py

@@ -2,23 +2,29 @@
 DAG to update daily notices metadata from TED.
 """
 
-from datetime import date, datetime
+from datetime import date, datetime, timedelta
 
-from airflow.models import Param
 from airflow.decorators import dag, task
+from airflow.models import Param
+from airflow.timetables.trigger import CronTriggerTimetable
 
 from dags import DEFAULT_DAG_ARGUMENTS
 from dags.dags_utils import get_dag_param
-from ted_sws.supra_notice_manager.services.daily_notices_metadata_services import update_daily_notices_metadata_from_ted, \
+from ted_sws.supra_notice_manager.services.daily_notices_metadata_services import \
+    update_daily_notices_metadata_from_ted, \
     update_daily_notices_metadata_with_fetched_data
 
 START_DATE_PARAM_KEY = "start_date"
 END_DATE_PARAM_KEY = "end_date"
+DEFAULT_TED_API_START_DATE = "2014-01-01"
+DEFAULT_TED_API_START_DATE_FORMAT = "%Y-%m-%d"
 
 
 @dag(default_args=DEFAULT_DAG_ARGUMENTS,
      schedule_interval=None,
      tags=['daily', "dashboards", "metadata", "ted", "notices"],
+     catchup=False,
+     timetable=CronTriggerTimetable('0 19 * * *', timezone='UTC'),
      description=__doc__[0: __doc__.find(".")],
      doc_md=__doc__,
      params={
@@ -43,19 +49,23 @@
 def daily_notices_metadata_update():
     @task
     def update_daily_notices_metadata_from_ted_api():
-        start_date = get_dag_param(key=START_DATE_PARAM_KEY, raise_error=True)
-        end_date = get_dag_param(key=END_DATE_PARAM_KEY, raise_error=True)
+        start_date = get_dag_param(key=START_DATE_PARAM_KEY, default_value=DEFAULT_TED_API_START_DATE)
+        end_date = get_dag_param(key=END_DATE_PARAM_KEY, default_value=(datetime.today() - timedelta(days=1)).strftime(
+            DEFAULT_TED_API_START_DATE_FORMAT))
 
-        update_daily_notices_metadata_from_ted(start_date=datetime.strptime(start_date, "%Y-%m-%d"),
-                                               end_date=datetime.strptime(end_date, "%Y-%m-%d"))
+        update_daily_notices_metadata_from_ted(
+            start_date=datetime.strptime(start_date, DEFAULT_TED_API_START_DATE_FORMAT),
+            end_date=datetime.strptime(end_date, DEFAULT_TED_API_START_DATE_FORMAT))
 
     @task
     def update_daily_notices_metadata_with_fetched_data_from_repo():
-        start_date = get_dag_param(key=START_DATE_PARAM_KEY, raise_error=True)
-        end_date = get_dag_param(key=END_DATE_PARAM_KEY, raise_error=True)
+        start_date = get_dag_param(key=START_DATE_PARAM_KEY, default_value=DEFAULT_TED_API_START_DATE)
+        end_date = get_dag_param(key=END_DATE_PARAM_KEY, default_value=(datetime.today() - timedelta(days=1)).strftime(
+            DEFAULT_TED_API_START_DATE_FORMAT))
 
-        update_daily_notices_metadata_with_fetched_data(start_date=datetime.strptime(start_date, "%Y-%m-%d"),
-                                                        end_date=datetime.strptime(end_date, "%Y-%m-%d"))
+        update_daily_notices_metadata_with_fetched_data(
+            start_date=datetime.strptime(start_date, DEFAULT_TED_API_START_DATE_FORMAT),
+            end_date=datetime.strptime(end_date, DEFAULT_TED_API_START_DATE_FORMAT))
 
     update_daily_notices_metadata_from_ted_api() >> update_daily_notices_metadata_with_fetched_data_from_repo()
 

ted_sws/supra_notice_manager/services/daily_notices_metadata_services.py

@@ -1,4 +1,4 @@
-from datetime import date, datetime, timedelta
+from datetime import date
 from typing import Optional, List
 
 from dateutil import rrule
@@ -11,8 +11,6 @@
 from ted_sws.data_manager.adapters.notice_repository import NoticeRepository
 from ted_sws.notice_fetcher.adapters.ted_api import TedAPIAdapter, TedRequestAPI
 
-DEFAULT_TED_API_START_DATE = "2014-01-01"
-DEFAULT_TED_API_START_DATE_FORMAT = "%Y-%m-%d"
 TED_API_NOTICE_ID_FIELD = "ND"
 TED_API_WILDCARD_DATE_FORMAT = "%Y%m%d*"
 DAILY_NOTICES_METADATA_TED_API_QUERY_RESULT_FIELDS = {"fields": ["ND"]}
@@ -57,20 +55,18 @@ def generate_list_of_dates_from_date_range(start_date: date, end_date: date) ->
     """
     if start_date > end_date:
         return None
-    return [ dt.date() for dt in rrule.rrule(rrule.DAILY,
-                                     dtstart=start_date,
-                                     until=end_date)]
+    return [dt.date() for dt in rrule.rrule(rrule.DAILY,
+                                            dtstart=start_date,
+                                            until=end_date)]
 
 
-def update_daily_notices_metadata_from_ted(start_date: date = None,
-                                           end_date: date = None,
+def update_daily_notices_metadata_from_ted(start_date: date,
+                                           end_date: date,
                                            ted_api: TedAPIAdapter = None,
                                            daily_notices_metadata_repo: DailyNoticesMetadataRepository = None):
     """
     Updates the daily notices metadata from the TED API.
     """
-    start_date = start_date or datetime.strptime(DEFAULT_TED_API_START_DATE, DEFAULT_TED_API_START_DATE_FORMAT)
-    end_date = end_date or datetime.today() - timedelta(days=1)
 
     if start_date > end_date:
         raise Exception("Start date cannot be greater than end date")
@@ -89,7 +85,7 @@ def update_daily_notices_metadata_from_ted(start_date: date = None,
 
     # Getting from TED API dates that are not in the repository from date range
     for day in dates_not_in_repository:
-        ted_api_query = DAILY_NOTICES_METADATA_TED_API_QUERY
+        ted_api_query = DAILY_NOTICES_METADATA_TED_API_QUERY.copy()
         ted_api_query[TED_API_QUERY_FIELD] = ted_api_query[TED_API_QUERY_FIELD].format(
             aggregation_date=day.strftime(TED_API_WILDCARD_DATE_FORMAT))
         notice_ids = ted_api.get_by_query(ted_api_query,
@@ -99,17 +95,14 @@ def update_daily_notices_metadata_from_ted(start_date: date = None,
         daily_notices_metadata_repo.add(daily_notices_metadata)
 
 
-def update_daily_notices_metadata_with_fetched_data(start_date: date = None,
-                                                    end_date: date = None,
+def update_daily_notices_metadata_with_fetched_data(start_date: date,
+                                                    end_date: date,
                                                     notice_repo: NoticeRepository = None,
                                                     daily_notices_metadata_repo: DailyNoticesMetadataRepository = None):
     """
     Updates the daily notices metadata witch fetched data.
     """
 
-    start_date = start_date or datetime.strptime(DEFAULT_TED_API_START_DATE, DEFAULT_TED_API_START_DATE_FORMAT)
-    end_date = end_date or datetime.today() - timedelta(days=1)
-
     if start_date > end_date:
         raise Exception("Start date cannot be greater than end date")
 
@@ -153,4 +146,3 @@ def update_daily_notices_metadata_with_fetched_data(start_date: date = None,
         daily_notices_metadata.mapping_suite_packages = mapping_suite_packages
         daily_notices_metadata.fetched_notice_ids = fetched_notice_ids
         daily_notices_metadata_repo.update(daily_notices_metadata)
-

tests/e2e/conftest.py

@@ -85,3 +85,8 @@ def fake_notice_repository(fake_mongodb_client):
 @pytest.fixture
 def example_date():
     return date(2021, 1, 7)
+
+
+@pytest.fixture
+def example_date_without_notices():
+    return date(2021, 1, 9)

tests/e2e/supra_notice_manager/test_notices_metadata_services.py

@@ -1,19 +1,27 @@
 from ted_sws import config
 from ted_sws.core.model.supra_notice import DailyNoticesMetadata
 from ted_sws.data_manager.adapters.daily_notices_metadata_repository import DailyNoticesMetadataRepository
-from ted_sws.supra_notice_manager.services.daily_notices_metadata_services import update_daily_notices_metadata_from_ted, \
-    update_daily_notices_metadata_with_fetched_data
 from ted_sws.notice_fetcher.adapters.ted_api import TedAPIAdapter, TedRequestAPI
+from ted_sws.supra_notice_manager.services.daily_notices_metadata_services import \
+    update_daily_notices_metadata_from_ted, \
+    update_daily_notices_metadata_with_fetched_data
 
 
-def test_update_daily_notices_metadata_from_ted(fake_mongodb_client, example_date):
+def test_update_daily_notices_metadata_from_ted(fake_mongodb_client, example_date, example_date_without_notices):
     """
     Test update_daily_notices_metadata_from_ted function
     """
 
     ted_api = TedAPIAdapter(TedRequestAPI(), config.TED_API_URL)
     daily_notices_metadata_repo = DailyNoticesMetadataRepository(fake_mongodb_client)
 
+    update_daily_notices_metadata_from_ted(start_date=example_date_without_notices,
+                                           end_date=example_date_without_notices,
+                                           ted_api=ted_api,
+                                           daily_notices_metadata_repo=daily_notices_metadata_repo)
+
+    assert len(list(daily_notices_metadata_repo.list())) == 1
+
     update_daily_notices_metadata_from_ted(start_date=example_date,
                                            end_date=example_date,
                                            ted_api=ted_api,
@@ -29,7 +37,7 @@ def test_update_daily_notices_metadata_from_ted(fake_mongodb_client, example_dat
                                            ted_api=ted_api,
                                            daily_notices_metadata_repo=daily_notices_metadata_repo)
 
-    assert len(list(daily_notices_metadata_repo.list())) == 1
+    assert len(list(daily_notices_metadata_repo.list())) == 2
 
 
 def test_update_daily_notices_metadata_with_fetched_data(fake_mongodb_client,