global: add pre-commit with ruff

.pre-commit-config.yaml

@@ -1,18 +1,19 @@
+files: ^backend/
 repos:
-  - repo: https://github.com/psf/black
-    rev: '24.2.0'
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.6.0
     hooks:
-      - id: black
-  - repo: https://github.com/pre-commit/mirrors-prettier
-    rev: 'v3.1.0'
+      - id: check-yaml
+      - id: end-of-file-fixer
+      - id: trailing-whitespace
+      - id: fix-byte-order-marker
+      - id: mixed-line-ending
+      - id: name-tests-test
+        args: [ --pytest-test-first ]
+        exclude: '^(?!factories/)'
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    rev: v0.5.6
     hooks:
-      - id: prettier
-  - repo: https://github.com/pycqa/isort
-    rev: '5.13.2'
-    hooks:
-      - id: isort
-  - repo: https://github.com/pycqa/flake8
-    rev: '7.0.0'
-    hooks:
-      - id: flake8
-        args: ['--config=backend/setup.cfg']
+      - id: ruff
+        args: [ --fix ]
+      - id: ruff-format

backend/gunicorn_config.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2020 CERN.
 #

backend/inspirehep/__init__.py

@@ -8,7 +8,6 @@
 
 """inspirehep."""
 
-
 from .version import __version__
 
 __all__ = ("__version__",)

backend/inspirehep/access_control.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/accounts/__init__.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/accounts/api.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #
@@ -84,8 +83,8 @@ def get_current_user_remote_orcid_account():
 
 
 def get_allowed_roles_for_collections(collections, read_only=True):
-    private_collections = set(collections) - (
-        current_app.config["NON_PRIVATE_LITERATURE_COLLECTIONS"]
+    private_collections = (
+        set(collections) - (current_app.config["NON_PRIVATE_LITERATURE_COLLECTIONS"])
     )
     roles = set()
     for col in private_collections:

backend/inspirehep/accounts/cli.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2023 CERN.
 #

backend/inspirehep/accounts/config.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2020 CERN.
 #
@@ -24,9 +23,9 @@
     "scope": " ".join(["/read-limited", "/activities/update", "/person/update"]),
     "show_login": "true",
 }
-REMOTE_SANDBOX_REST_APP["signup_handler"][
-    "setup"
-] = "inspirehep.accounts.handlers:account_setup_handler"
+REMOTE_SANDBOX_REST_APP["signup_handler"]["setup"] = (
+    "inspirehep.accounts.handlers:account_setup_handler"
+)
 # ==============================================================================
 
 # ORCID production =============================================================
@@ -37,9 +36,9 @@
     "scope": " ".join(["/read-limited", "/activities/update", "/person/update"]),
     "show_login": "true",
 }
-REMOTE_REST_APP["signup_handler"][
-    "setup"
-] = "inspirehep.accounts.handlers:account_setup_handler"
+REMOTE_REST_APP["signup_handler"]["setup"] = (
+    "inspirehep.accounts.handlers:account_setup_handler"
+)
 REMOTE_REST_APP["params"]["access_token_url"] = "https://api.orcid.org/oauth/token"
 REMOTE_REST_APP["params"]["authorize_url"] = "https://orcid.org/oauth/authorize"
 # ==============================================================================

backend/inspirehep/accounts/decorators.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/accounts/fixtures.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/accounts/handlers.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2020 CERN.
 #
@@ -12,9 +11,10 @@
 from invenio_oauthclient.utils import oauth_link_external_id
 from invenio_pidstore.models import PersistentIdentifier
 
-from .api import get_current_user_remote_orcid_account
+from inspirehep.accounts.api import get_current_user_remote_orcid_account
 from inspirehep.records.utils import get_pid_for_pid
 
+
 def get_current_user_data():
     remote_orcid_account = get_current_user_remote_orcid_account()
     orcid_account_extra_data = (
@@ -26,7 +26,7 @@ def get_current_user_data():
             "email": current_user.email,
             "roles": [role.name for role in current_user.roles],
             "orcid": orcid,
-            "profile_control_number": get_pid_for_pid('orcid', orcid, 'recid'),
+            "profile_control_number": get_pid_for_pid("orcid", orcid, "recid"),
             "allow_orcid_push": orcid_account_extra_data.get("allow_push"),
         }
     }

backend/inspirehep/accounts/roles.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/accounts/views.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2020 CERN.
 #
@@ -15,13 +14,12 @@
 from webargs import fields
 from webargs.flaskparser import FlaskParser
 
+from inspirehep.accounts.api import get_current_user_remote_orcid_account
+from inspirehep.accounts.decorators import login_required
+from inspirehep.accounts.handlers import get_current_user_data
 from inspirehep.orcid.tasks import push_account_literature_to_orcid
 from inspirehep.serializers import jsonify
 
-from .api import get_current_user_remote_orcid_account
-from .decorators import login_required
-from .handlers import get_current_user_data
-
 parser = FlaskParser()
 
 blueprint = Blueprint(

backend/inspirehep/alembic/35ba3d715114_update_primary_key_in_students_advisors.py

@@ -5,8 +5,7 @@
 # Invenio is free software; you can redistribute it and/or modify it
 # under the terms of the MIT License; see LICENSE file for more details.
 
-"""update primary key in students advisors
-"""
+"""update primary key in students advisors"""
 
 import sqlalchemy as sa
 from alembic import op

backend/inspirehep/alembic/595c36d68964_add_authors_to_records_table_and_update_.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #
@@ -9,11 +8,10 @@
 
 import sqlalchemy as sa
 from alembic import op
+from inspirehep.records.models import AuthorSchemaType
 from sqlalchemy.dialects.postgresql import ENUM
 from sqlalchemy_utils import UUIDType
 
-from inspirehep.records.models import AuthorSchemaType
-
 revision = "595c36d68964"
 down_revision = "cea5fa2e5d2c"
 branch_labels = ()

backend/inspirehep/alembic/7be4c8b5c5e8_inspirehep_new_migrations.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #
@@ -8,7 +7,6 @@
 """Inspirehep initial revision of migrations
 which makes db identical like in inspire-next"""
 
-
 # revision identifiers, used by Alembic.
 revision = "7be4c8b5c5e8"
 down_revision = "b5be5fda2ee7"

backend/inspirehep/alembic/8ba47044154a_add_indexed_to_records_citations_on_is_.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/alembic/b495825c322b_add_recid_to_recordsauthors_id_type.py

@@ -6,6 +6,7 @@
 # under the terms of the MIT License; see LICENSE file for more details.
 
 """add recid to RecordsAuthors id_type"""
+
 from alembic import op
 
 # revision identifiers, used by Alembic.

backend/inspirehep/alembic/c6570e49b7b2_change_record_citations_to_records_.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/alembic/e5e43ad8f861_add_index_for_objectversion_key.py

@@ -7,7 +7,6 @@
 
 """Add index for ObjectVersion.key"""
 
-
 # revision identifiers, used by Alembic.
 revision = "e5e43ad8f861"
 down_revision = "788a3a61a635"

backend/inspirehep/assign/tasks.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2021 CERN.
 #
@@ -117,9 +116,9 @@ def assign_conference(record, conference_ref, cnum):
                     cnum=cnum,
                 )
             elif cnum == record_cnum:
-                builder.record["publication_info"][idx][
-                    "conference_record"
-                ] = conference_ref
+                builder.record["publication_info"][idx]["conference_record"] = (
+                    conference_ref
+                )
                 LOGGER.warning(
                     "conference cnum already assigned to paper without ref.",
                     recid=record.control_number,
@@ -226,9 +225,9 @@ def create_rt_ticket_for_claiming_action(
             from_author_recid, lit_record.get("authors", [])
         )
         if matched_author_name:
-            incompatibile_names_papers[
-                f"{INSPIREHEP_URL}/literature/{paper_recid}"
-            ] = matched_author_name
+            incompatibile_names_papers[f"{INSPIREHEP_URL}/literature/{paper_recid}"] = (
+                matched_author_name
+            )
 
     to_author_names = []
     to_author_names.extend(get_value(to_author_record, "name.name_variants", []))

backend/inspirehep/assign/utils.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #
@@ -117,10 +116,7 @@ def get_last_names(name):
             return {unidecode(parsed_name.first)}
         # corner case for full names without comma,
         # we are treating them as last names (ie. "Smith Davis")
-        if "," not in name:
-            names = name.split()
-        else:
-            names = parsed_name.last_list
+        names = name.split() if "," not in name else parsed_name.last_list
 
         return {unidecode(name) for name in names}
 
@@ -148,7 +144,7 @@ def _check_names_compability(lit_record, author_parsed_name, last_names_only=Fal
     )
     matched_authors_recids = [
         recid
-        for recid in lit_authors_names_recids.keys()
+        for recid in lit_authors_names_recids
         if lit_authors_names_recids[recid] == author_name_to_compare
     ]
     if len(matched_authors_recids) == 1:

backend/inspirehep/assign/views.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2020 CERN.
 #

backend/inspirehep/cds/api.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2021 CERN.
 #
@@ -8,13 +7,13 @@
 
 import requests
 import structlog
-from opensearch_dsl import Q
 from flask import current_app
 from idutils import is_arxiv
 from inspire_schemas.builders import LiteratureBuilder
 from inspire_utils.record import get_value, get_values_for_schema
 from invenio_db import db
 from invenio_pidstore.errors import PIDDoesNotExistError
+from opensearch_dsl import Q
 from sqlalchemy.exc import SQLAlchemyError
 
 from inspirehep.cds.errors import CDSSyncError, MissingCDSServerConfig

backend/inspirehep/cds/cli.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2021 CERN.
 #
@@ -42,10 +41,10 @@ def sync(since=None):
     elif not isinstance(since, date):
         try:
             since = datetime.strptime(since, "%Y-%m-%d").date()
-        except ValueError:
+        except ValueError as e:
             raise WrongDateFormat(
                 f"`since`: {since} is in wrong format. Should be in ISO format: YYYY-MM-DD."
-            )
+            ) from e
     try:
         click.echo("Starting CDS Sync.")
         sync_identifiers(since)

backend/inspirehep/cds/config.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2021 CERN.
 #

backend/inspirehep/cds/errors.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2021 CERN.
 #

backend/inspirehep/cds/models.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/celery.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #
@@ -26,7 +25,7 @@
 class CeleryTask(AppContextTask):
     def on_failure(self, exc, task_id, args, kwargs, einfo):
         if isinstance(
-            exc, (InvalidRequestError, OperationalError, Psycopg2OperationalError)
+            exc, InvalidRequestError | OperationalError | Psycopg2OperationalError
         ):
             LOGGER.exception("Shutting down celery process")
             try:

backend/inspirehep/config.py

@@ -1,4 +1,3 @@
-# -*- coding: utf-8 -*-
 #
 # Copyright (C) 2019 CERN.
 #

backend/inspirehep/curation/__init__.py

@@ -0,0 +1,5 @@
+#
+# Copyright (C) 2021 CERN.
+#
+# inspirehep is free software; you can redistribute it and/or modify it under
+# the terms of the MIT License; see LICENSE file for more details.