global: add pre-commit with ruff

.pre-commit-config.yaml

@@ -0,0 +1,17 @@
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.6.0
+    hooks:
+      - 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: ruff
+        args: [ --fix ]

inspire_dojson/api.py

@@ -28,16 +28,15 @@
 import re
 from itertools import chain
 
+from dojson.contrib.marc21.utils import create_record
+from inspire_utils.helpers import force_list
+from inspire_utils.record import get_value
 from lxml.builder import E
 from lxml.etree import tostring
 from six import iteritems, text_type, unichr
 from six.moves import urllib
 
-from dojson.contrib.marc21.utils import create_record
-
 from inspire_dojson.utils import create_record_from_dict, force_single_element
-from inspire_utils.helpers import force_list
-from inspire_utils.record import get_value
 
 from .cds import cds2hep_marc
 from .conferences import conferences

inspire_dojson/cds/model.py

@@ -25,8 +25,9 @@
 from __future__ import absolute_import, division, print_function
 
 from itertools import chain
-from inspire_utils.record import get_value
+
 from inspire_utils.helpers import force_list
+from inspire_utils.record import get_value
 
 from ..model import FilterOverdo, clean_record
 

inspire_dojson/cds/rules.py

@@ -31,17 +31,14 @@
 import pycountry
 import rfc3987
 import six
-
-from idutils import is_arxiv
 from dojson import utils
-
-from six.moves import urllib
-
+from idutils import is_arxiv
 from inspire_utils.helpers import force_list
 from inspire_utils.name import normalize_name
+from six.moves import urllib
 
-from .model import cds2hep_marc
 from ..utils import force_single_element, quote_url
+from .model import cds2hep_marc
 
 CATEGORIES = {
     'Accelerators and Storage Rings': 'Accelerators',

inspire_dojson/common/rules.py

@@ -27,15 +27,13 @@
 import re
 from datetime import datetime
 
-from flask import current_app
-from six.moves import urllib
-
 from dojson import utils
-
+from flask import current_app
 from inspire_schemas.api import load_schema
 from inspire_schemas.utils import classify_field
 from inspire_utils.date import PartialDate, earliest_date
 from inspire_utils.helpers import force_list, maybe_int
+from six.moves import urllib
 
 from ..conferences.model import conferences
 from ..data.model import data
@@ -50,7 +48,6 @@
     get_record_ref,
 )
 
-
 IS_INTERNAL_UID = re.compile(r'^(inspire:uid:)?\d{5}$')
 IS_ORCID = re.compile(r'^(orcid:)?\d{4}-\d{4}-\d{4}-\d{3}[0-9X]$')
 

inspire_dojson/conferences/model.py

@@ -24,7 +24,7 @@
 
 from __future__ import absolute_import, division, print_function
 
-from ..model import FilterOverdo, add_schema, add_collection, clean_record
+from ..model import FilterOverdo, add_collection, add_schema, clean_record
 
 
 def remove_lone_series_number(record, blob):

inspire_dojson/conferences/rules.py

@@ -25,12 +25,11 @@
 from __future__ import absolute_import, division, print_function
 
 from dojson import utils
+from inspire_utils.helpers import force_list, maybe_float, maybe_int
 
-from inspire_utils.helpers import force_list, maybe_int, maybe_float
-
-from .model import conferences
 from ..utils import force_single_element
 from ..utils.geo import parse_conference_address
+from .model import conferences
 
 
 def _trim_date(date):

inspire_dojson/data/model.py

@@ -24,8 +24,7 @@
 
 from __future__ import absolute_import, division, print_function
 
-from ..model import FilterOverdo, add_schema, add_collection, clean_record
-
+from ..model import FilterOverdo, add_collection, add_schema, clean_record
 
 filters = [
     add_schema('data.json'),

inspire_dojson/data/rules.py

@@ -27,8 +27,8 @@
 from dojson import utils
 from idutils import normalize_doi
 
-from .model import data
 from ..utils import force_single_element, get_record_ref
+from .model import data
 
 
 @data.over('dois', '^0247.')

inspire_dojson/experiments/rules.py

@@ -26,13 +26,11 @@
 
 from dojson import utils
 from dojson.errors import IgnoreKey
-
 from inspire_utils.date import normalize_date
 from inspire_utils.helpers import force_list, maybe_int
 
-from .model import experiments
 from ..utils import force_single_element, get_record_ref
-
+from .model import experiments
 
 EXPERIMENT_CATEGORIES_MAP = \
     {'1': 'Collider Experiments',

inspire_dojson/hep/__init__.py

@@ -24,6 +24,7 @@
 
 from __future__ import absolute_import, division, print_function
 
+from .model import hep, hep2marc  # noqa: F401
 from .rules import (  # noqa: F401
     bd0xx,
     bd1xx,
@@ -36,4 +37,3 @@
     bd9xx,
     bdFFT,
 )
-from .model import hep, hep2marc  # noqa: F401

inspire_dojson/hep/model.py

@@ -27,12 +27,11 @@
 import itertools
 
 import six
-
+from inspire_schemas.builders.literature import is_citeable
 from inspire_schemas.utils import (
     convert_old_publication_info_to_new,
     normalize_arxiv_category,
 )
-from inspire_schemas.builders.literature import is_citeable
 from inspire_utils.helpers import force_list
 from inspire_utils.record import get_value
 

inspire_dojson/hep/rules/bd0xx.py

@@ -28,17 +28,14 @@
 from collections import defaultdict
 
 import pycountry
-
 from dojson import utils
 from idutils import is_arxiv_post_2007, is_doi, is_handle, normalize_doi
-
 from inspire_schemas.api import load_schema
 from inspire_schemas.utils import normalize_arxiv_category
 from inspire_utils.helpers import force_list
 
-from ..model import hep, hep2marc
 from ...utils import force_single_element, normalize_isbn
-
+from ..model import hep, hep2marc
 
 RE_LANGUAGE = re.compile(r'\/| or | and |,|=|\s+')
 

inspire_dojson/hep/rules/bd1xx.py

@@ -27,16 +27,14 @@
 import re
 
 from dojson import utils
-
 from inspire_utils.dedupers import dedupe_list
 from inspire_utils.helpers import force_list, maybe_int
 
-from ..model import hep, hep2marc
 from ...utils import (
     force_single_element,
     get_record_ref,
 )
-
+from ..model import hep, hep2marc
 
 ORCID = re.compile(r'\d{4}-\d{4}-\d{4}-\d{3}[0-9Xx]')
 

inspire_dojson/hep/rules/bd2xx.py

@@ -25,13 +25,11 @@
 from __future__ import absolute_import, division, print_function
 
 import langdetect
-
 from dojson import utils
-
 from inspire_utils.helpers import force_list
 
-from ..model import hep, hep2marc
 from ...utils import normalize_date_aggressively
+from ..model import hep, hep2marc
 
 
 @hep.over('titles', '^(210|245|246|247)..')

inspire_dojson/hep/rules/bd3xx.py

@@ -26,8 +26,8 @@
 
 from inspire_utils.helpers import maybe_int
 
-from ..model import hep, hep2marc
 from ...utils import force_single_element
+from ..model import hep, hep2marc
 
 
 @hep.over('number_of_pages', '^300..')

inspire_dojson/hep/rules/bd5xx.py

@@ -27,13 +27,11 @@
 import re
 
 from dojson import utils
-
 from inspire_utils.date import normalize_date
 from inspire_utils.helpers import force_list, maybe_int
 
-from ..model import hep, hep2marc
 from ...utils import force_single_element, get_record_ref
-
+from ..model import hep, hep2marc
 
 IS_DEFENSE_DATE = re.compile('Presented (on )?(?P<defense_date>.*)', re.IGNORECASE)
 

inspire_dojson/hep/rules/bd6xx.py

@@ -25,13 +25,11 @@
 from __future__ import absolute_import, division, print_function
 
 import six
-
 from dojson import utils
-
 from inspire_utils.helpers import force_list
 
-from ..model import hep, hep2marc
 from ...utils import force_single_element, get_record_ref
+from ..model import hep, hep2marc
 
 ENERGY_RANGES_MAP = {
     '1': '0-3 GeV',

inspire_dojson/hep/rules/bd7xx.py

@@ -25,7 +25,6 @@
 from __future__ import absolute_import, division, print_function
 
 from dojson import utils
-
 from inspire_schemas.api import load_schema
 from inspire_schemas.utils import (
     convert_new_publication_info_to_old,
@@ -34,13 +33,13 @@
 )
 from inspire_utils.helpers import force_list, maybe_int
 
-from ..model import hep, hep2marc
 from ...utils import (
     force_single_element,
     get_recid_from_ref,
     get_record_ref,
     normalize_isbn,
 )
+from ..model import hep, hep2marc
 
 
 @hep.over('collaborations', '^710..')

inspire_dojson/hep/rules/bd9xx.py

@@ -28,7 +28,6 @@
 
 from dojson import utils
 from idutils import is_arxiv_post_2007
-
 from inspire_schemas.api import ReferenceBuilder, load_schema
 from inspire_schemas.utils import (
     build_pubnote,
@@ -38,8 +37,8 @@
 from inspire_utils.helpers import force_list, maybe_int
 from inspire_utils.record import get_value
 
-from ..model import hep, hep2marc
 from ...utils import force_single_element, get_recid_from_ref, get_record_ref
+from ..model import hep, hep2marc
 
 COLLECTIONS_MAP = {
     'babar-analysisdocument': 'BABAR Analysis Documents',

inspire_dojson/hep/rules/bdFFT.py

@@ -28,11 +28,10 @@
 import re
 
 from dojson import utils
-
 from inspire_utils.helpers import force_list
 
-from ..model import hep, hep2marc
 from ...utils import absolute_url, afs_url, afs_url_to_path
+from ..model import hep, hep2marc
 
 
 @hep.over('documents', '^FFT[^%][^%]')

inspire_dojson/hepnames/model.py

@@ -24,8 +24,7 @@
 
 from __future__ import absolute_import, division, print_function
 
-from ..model import FilterOverdo, add_schema, add_collection, clean_marc, clean_record
-
+from ..model import FilterOverdo, add_collection, add_schema, clean_marc, clean_record
 
 hepnames_filters = [
     add_schema('authors.json'),

inspire_dojson/hepnames/rules.py

@@ -27,7 +27,6 @@
 import re
 
 from dojson import utils
-
 from inspire_schemas.api import load_schema
 from inspire_schemas.utils import (
     normalize_arxiv_category,
@@ -37,16 +36,15 @@
 from inspire_utils.helpers import force_list, maybe_int
 from inspire_utils.name import normalize_name
 
-from .model import hepnames, hepnames2marc
 from ..utils import (
     force_single_element,
-    get_record_ref,
     get_recid_from_ref,
+    get_record_ref,
     normalize_rank,
     quote_url,
-    unquote_url
+    unquote_url,
 )
-
+from .model import hepnames, hepnames2marc
 
 AWARD_YEAR = re.compile(r'\(?(?P<year>\d{4})\)?')
 INSPIRE_BAI = re.compile(r'(\w+\.)+\d+')
@@ -760,7 +758,7 @@ def _is_deleted(value):
         return force_single_element(value.get('c', '')).upper() == 'DELETED'
 
     def _is_stub(value):
-        return not (force_single_element(value.get('a', '')).upper() == 'USEFUL')
+        return force_single_element(value.get('a', '')).upper() != 'USEFUL'
 
     deleted = self.get('deleted')
     stub = self.get('stub')

inspire_dojson/institutions/model.py

@@ -24,7 +24,7 @@
 
 from __future__ import absolute_import, division, print_function
 
-from ..model import FilterOverdo, add_schema, add_collection, clean_record
+from ..model import FilterOverdo, add_collection, add_schema, clean_record
 
 
 def combine_addresses_and_location(record, blob):

inspire_dojson/institutions/rules.py

@@ -27,13 +27,11 @@
 import re
 
 from dojson import utils
-
 from inspire_utils.helpers import force_list, maybe_float, maybe_int
 
-from .model import institutions
 from ..utils import force_single_element, get_record_ref
 from ..utils.geo import parse_institution_address
-
+from .model import institutions
 
 ACRONYM = re.compile(r'\s*\((.*)\)\s*$')