Update workflows (#3858)

* Update workflows

* Apply ruff-format updates

---------

Co-authored-by: Krystle Salazar <[email protected]>

.github/workflows/pr_label_check.yml

@@ -42,7 +42,7 @@ jobs:
       - get_label_groups
     steps:
       - name: Check aspect label
-        uses: docker://agilepathway/pull-request-label-checker:v1.6.20
+        uses: docker://agilepathway/pull-request-label-checker:v1.6.23
         with:
           any_of: ${{ needs.get_label_groups.outputs.aspect }}
           repo_token: ${{ secrets.GITHUB_TOKEN }}
@@ -55,7 +55,7 @@ jobs:
       - get_label_groups
     steps:
       - name: Check goal label
-        uses: docker://agilepathway/pull-request-label-checker:v1.6.20
+        uses: docker://agilepathway/pull-request-label-checker:v1.6.23
         with:
           one_of: ${{ needs.get_label_groups.outputs.goal }}
           repo_token: ${{ secrets.GITHUB_TOKEN }}
@@ -68,7 +68,7 @@ jobs:
       - get_label_groups
     steps:
       - name: Check priority label
-        uses: docker://agilepathway/pull-request-label-checker:v1.6.20
+        uses: docker://agilepathway/pull-request-label-checker:v1.6.23
         with:
           one_of: ${{ needs.get_label_groups.outputs.priority }}
           repo_token: ${{ secrets.GITHUB_TOKEN }}
@@ -81,7 +81,7 @@ jobs:
       - get_label_groups
     steps:
       - name: Check stack label
-        uses: docker://agilepathway/pull-request-label-checker:v1.6.20
+        uses: docker://agilepathway/pull-request-label-checker:v1.6.23
         with:
           any_of: ${{ needs.get_label_groups.outputs.stack }}
           repo_token: ${{ secrets.GITHUB_TOKEN }}

.pre-commit-config.yaml

@@ -46,7 +46,7 @@ repos:
       - id: requirements-txt-fixer
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.2.0
+    rev: v0.3.0
     hooks:
       - id: ruff # replaces Flake8, isort, pydocstyle, pyupgrade
         args:
@@ -74,7 +74,7 @@ repos:
       - id: shfmt-docker
 
   - repo: https://github.com/rhysd/actionlint
-    rev: v1.6.26
+    rev: v1.6.27
     hooks:
       - id: actionlint-docker
 
@@ -153,7 +153,7 @@ repos:
         files: (.vale/.*|.mdx?)$
 
   - repo: https://github.com/renovatebot/pre-commit-hooks
-    rev: 37.168.1
+    rev: 37.221.1
     hooks:
       - id: renovate-config-validator
         args:

api/api/constants/media_types.py

@@ -1,4 +1,5 @@
 """Also see `ingestion_server/constants/media_types.py`."""
+
 from typing import Literal
 
 

api/api/utils/aiohttp.py

@@ -14,9 +14,9 @@
     asyncio.AbstractEventLoop, aiohttp.ClientSession
 ] = weakref.WeakKeyDictionary()
 
-_LOCKS: weakref.WeakKeyDictionary[
-    asyncio.AbstractEventLoop, asyncio.Lock
-] = weakref.WeakKeyDictionary()
+_LOCKS: weakref.WeakKeyDictionary[asyncio.AbstractEventLoop, asyncio.Lock] = (
+    weakref.WeakKeyDictionary()
+)
 
 
 @asgi_shutdown.connect

api/api/utils/licenses.py

@@ -5,7 +5,6 @@
 frontend, or open an issue to track it.
 """
 
-
 from api.constants.licenses import (
     ALL_CC_LICENSES,
     DEPRECATED_CC_LICENSES,

api/conf/urls/__init__.py

@@ -4,6 +4,7 @@
 For more information on this file, see
 https://docs.djangoproject.com/en/4.2/topics/http/urls/
 """
+
 from django.conf import settings
 from django.conf.urls.static import static
 from django.contrib import admin

api/test/fixtures/asynchronous.py

@@ -43,8 +43,7 @@ def ensure_asgi_lifecycle(session_loop: asyncio.AbstractEventLoop):
     """
     scope = {"type": "lifespan"}
 
-    async def noop(*args, **kwargs):
-        ...
+    async def noop(*args, **kwargs): ...
 
     async def shutdown():
         return {"type": "lifespan.shutdown"}

api/test/unit/utils/test_throttle.py

@@ -43,9 +43,9 @@ def enable_throttles(settings):
     # Put settings into base Django settings from which DRF reads
     # settings when we call `api_settings.reload()`
     settings.REST_FRAMEWORK["DEFAULT_THROTTLE_RATES"] = settings.DEFAULT_THROTTLE_RATES
-    settings.REST_FRAMEWORK[
-        "DEFAULT_THROTTLE_CLASSES"
-    ] = settings.DEFAULT_THROTTLE_CLASSES
+    settings.REST_FRAMEWORK["DEFAULT_THROTTLE_CLASSES"] = (
+        settings.DEFAULT_THROTTLE_CLASSES
+    )
 
     # Reload the settings and read them from base Django settings
     # Also handles importing classes from class strings, etc

automations/python/models/label.py

@@ -4,7 +4,6 @@
 
 
 class Label:
-
     """
     Represents a single label.
 

automations/python/workflows/get_folder_differences.py

@@ -7,6 +7,7 @@
 files on the docs preview site. Finally, it outputs the preview site docs
 generated comment along with a list of new and changed files.
 """
+
 import os
 import shutil
 import subprocess

catalog/dags/common/cloudwatch.py

@@ -9,6 +9,7 @@
 in the Airflow Variables, which is particularly the desired behavior when running
 the Data Refresh DAGs locally or in a development environment.
 """
+
 import logging
 
 from airflow.exceptions import AirflowSkipException

catalog/dags/common/licenses/licenses.py

@@ -2,6 +2,7 @@
 This module has a number of public methods which are useful for working
 with licenses.
 """
+
 import logging
 from functools import lru_cache
 from typing import NamedTuple

catalog/dags/common/loader/provider_details.py

@@ -8,6 +8,7 @@
 might be useful for retrieving sub-providers at the database level and the
 API level.
 """
+
 import os
 
 

catalog/dags/common/storage/db_columns.py

@@ -2,6 +2,7 @@
 This module contains the lists of database columns in the same order as in the
 main media tables within the database.
 """
+
 from common.constants import AUDIO, IMAGE
 from common.storage import columns as col
 from common.utils import setup_kwargs_for_media_type

catalog/dags/common/storage/util.py

@@ -1,4 +1,5 @@
 """This module has public methods which are useful for storage operations."""
+
 import logging
 
 from common.storage.audio import AudioStore

catalog/dags/common/urls.py

@@ -2,6 +2,7 @@
 This module has a number of public methods which are useful for
 verifying and cleaning URLs.
 """
+
 import logging
 import re
 from functools import lru_cache

catalog/dags/data_refresh/create_filtered_index.py

@@ -15,6 +15,7 @@
 finally deleting the old, now unused filtered index. These TaskGroups are used in
 the data refresh DAGs to execute the filtered index steps.
 """
+
 from datetime import timedelta
 
 from airflow.operators.empty import EmptyOperator

catalog/dags/data_refresh/create_filtered_index_dag.py

@@ -50,6 +50,7 @@
 This ensures that neither are depending on or modifying the origin indexes
 critical for the creation of the filtered indexes.
 """
+
 from datetime import datetime
 
 from airflow import DAG

catalog/dags/data_refresh/dag_factory.py

@@ -22,6 +22,7 @@
 - [[Feature] Merge popularity calculations and data refresh into a single DAG](
 https://github.com/WordPress/openverse-catalog/issues/453)
 """
+
 import logging
 from collections.abc import Sequence
 

catalog/dags/data_refresh/data_refresh_task_factory.py

@@ -44,6 +44,7 @@
 - [[Feature] Data refresh orchestration DAG](
 https://github.com/WordPress/openverse-catalog/issues/353)
 """
+
 import logging
 import os
 from collections.abc import Sequence

catalog/dags/data_refresh/data_refresh_types.py

@@ -5,6 +5,7 @@
 for each of our media types. This configuration information is used to generate
 the dynamic Data Refresh dags.
 """
+
 import os
 from dataclasses import dataclass, field
 from datetime import datetime, timedelta

catalog/dags/database/batched_update/batched_update_dag.py

@@ -63,7 +63,6 @@
 failures during the `update_batches` step.
 """
 
-
 import logging
 
 from airflow.decorators import dag

catalog/dags/database/delete_records/delete_records_dag.py

@@ -50,7 +50,6 @@
 DAGs are not currently running.
 """
 
-
 import logging
 
 from airflow.decorators import dag

catalog/dags/database/report_pending_reported_media.py

@@ -9,6 +9,7 @@
 taken. If a record has been reported multiple times, it only needs to be
 reviewed once and so is only counted once in the reporting by this DAG.
 """
+
 import logging
 import os
 from textwrap import dedent

catalog/dags/elasticsearch_cluster/create_new_es_index/create_new_es_index_dag.py

@@ -96,6 +96,7 @@
 }
 ```
 """
+
 import logging
 
 from airflow import DAG

catalog/dags/elasticsearch_cluster/create_proportional_by_source_staging_index/create_proportional_by_source_staging_index_dag.py

@@ -36,6 +36,7 @@
 * `recreate_full_staging_index`
 * `create_new_staging_es_index`
 """
+
 from datetime import datetime, timedelta
 
 from airflow.decorators import dag

catalog/dags/elasticsearch_cluster/recreate_staging_index/recreate_full_staging_index_dag.py

@@ -40,6 +40,7 @@
 * `create_proportional_by_provider_staging_index`
 * `create_new_staging_es_index`
 """
+
 from datetime import datetime
 
 from airflow.decorators import dag

catalog/dags/flickr_thumbs_removal.py

@@ -2,6 +2,7 @@
 One-time run DAG to remove progressively all the old Flickr thumbnails,
 as they were determined to be unsuitable for the Openverse UI requirements.
 """
+
 import logging
 from datetime import timedelta
 from textwrap import dedent

catalog/dags/maintenance/add_license_url.py

@@ -9,6 +9,7 @@
 the `meta_data` column are updated, the DAG will only run the first and the
 last step, logging the statistics.
 """
+
 import csv
 import logging
 from collections import defaultdict

catalog/dags/maintenance/airflow_log_cleanup_workflow.py

@@ -19,6 +19,7 @@
 - maxLogAgeInDays:<INT> - Optional
 - enableDelete:<BOOLEAN> - Optional
 """
+
 from datetime import datetime, timedelta
 
 import jinja2

catalog/dags/oauth2/authorize_dag.py

@@ -1,4 +1,5 @@
 """# OAuth Provider Authorization"""
+
 from datetime import datetime
 
 from airflow.models import DAG

catalog/dags/oauth2/token_refresh_dag.py

@@ -1,4 +1,5 @@
 """# OAuth Provider Token Refresh"""
+
 from datetime import datetime
 
 from airflow.models import DAG

catalog/dags/popularity/popularity_refresh_dag_factory.py

@@ -22,6 +22,7 @@
 - [[Implementation Plan] Decoupling Popularity Calculations from the Data Refresh](
 https://docs.openverse.org/projects/proposals/popularity_optimizations/20230420-implementation_plan_popularity_optimizations.html)
 """
+
 import logging
 from datetime import datetime
 

catalog/dags/popularity/popularity_refresh_types.py

@@ -5,6 +5,7 @@
 `POPULARITY_REFRESH_CONFIGS` for each of our media types. This configuration info
 is used to generate the dynamic Popularity Refresh dags.
 """
+
 import os
 from dataclasses import dataclass, field
 from datetime import datetime, timedelta

catalog/dags/popularity/recreate_popularity_calculation_dag_factory.py

@@ -12,6 +12,7 @@
 These DAGs are not on a schedule, and should only be run manually when new
 SQL code is deployed for the calculation.
 """
+
 from airflow import DAG
 
 from common.constants import DAG_DEFAULT_ARGS, POSTGRES_CONN_ID
@@ -74,6 +75,6 @@ def create_recreate_popularity_calculation_dag(popularity_refresh: PopularityRef
     recreate_popularity_calculation_dag = create_recreate_popularity_calculation_dag(
         popularity_refresh
     )
-    globals()[
-        recreate_popularity_calculation_dag.dag_id
-    ] = recreate_popularity_calculation_dag
+    globals()[recreate_popularity_calculation_dag.dag_id] = (
+        recreate_popularity_calculation_dag
+    )

catalog/dags/providers/provider_api_scripts/auckland_museum.py

@@ -16,6 +16,7 @@
 /search, /id | 10 | 1000
 /id/media | 10 | 1000
 """
+
 import logging
 from datetime import datetime, timedelta
 

catalog/dags/providers/provider_api_scripts/europeana.py

@@ -8,6 +8,7 @@
 
 Notes:                  https://pro.europeana.eu/page/search
 """
+
 import argparse
 import functools
 import logging

catalog/dags/providers/provider_api_scripts/finnish_museums.py

@@ -14,6 +14,7 @@
                         reingestion DAG, as updated data will be processed during
                         regular ingestion.
 """
+
 import logging
 from itertools import chain
 

catalog/dags/providers/provider_api_scripts/freesound.py

@@ -11,6 +11,7 @@
                         This script can be run either to ingest the full dataset or
                         as a dated DAG.
 """
+
 import functools
 import logging
 from datetime import datetime

catalog/dags/providers/provider_api_scripts/jamendo.py

@@ -15,6 +15,7 @@
                         sample rate: 44.1 or 48 kHz
                         channels: 1/2
 """
+
 import logging
 from datetime import timedelta
 from urllib.parse import parse_qs, urlencode, urlsplit

catalog/dags/providers/provider_api_scripts/justtakeitfree.py

@@ -9,6 +9,7 @@
 Notes:                  https://justtakeitfree.com/api/api.php
 This API requires an API key. For more details, see https://github.com/WordPress/openverse/pull/2793
 """
+
 import logging
 
 from airflow.models import Variable

catalog/dags/providers/provider_api_scripts/nappy.py

@@ -10,6 +10,7 @@
                         https://nappy.co/
 
 """
+
 import logging
 
 from common import constants