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, --unsafe-fixes ]

inspire_dojson/__init__.py

@@ -24,8 +24,8 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import common  # noqa: F401
-from .api import marcxml2record, record2marcxml  # noqa: F401
-from .errors import DoJsonError  # noqa: F401
+from inspire_dojson import common  # noqa: F401
+from inspire_dojson.api import marcxml2record, record2marcxml  # noqa: F401
+from inspire_dojson.errors import DoJsonError  # noqa: F401
 
 __version__ = "63.2.22"

inspire_dojson/api.py

@@ -28,26 +28,24 @@
 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.cds import cds2hep_marc
+from inspire_dojson.conferences import conferences
+from inspire_dojson.data import data
+from inspire_dojson.errors import NotSupportedError
+from inspire_dojson.experiments import experiments
+from inspire_dojson.hep import hep, hep2marc
+from inspire_dojson.hepnames import hepnames, hepnames2marc
+from inspire_dojson.institutions import institutions
+from inspire_dojson.journals import journals
 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
-from .data import data
-from .errors import NotSupportedError
-from .experiments import experiments
-from .hep import hep, hep2marc
-from .hepnames import hepnames, hepnames2marc
-from .institutions import institutions
-from .journals import journals
 
 try:
     unichr(0x100000)

inspire_dojson/cds/__init__.py

@@ -24,5 +24,5 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import rules  # noqa: F401
-from .model import cds2hep_marc  # noqa: F401
+from inspire_dojson.cds import rules  # noqa: F401
+from inspire_dojson.cds.model import cds2hep_marc  # noqa: F401

inspire_dojson/cds/model.py

@@ -25,10 +25,11 @@
 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
+from inspire_dojson.model import FilterOverdo, clean_record
 
 
 def add_control_number(record, blob):

inspire_dojson/cds/rules.py

@@ -24,24 +24,22 @@
 
 from __future__ import absolute_import, division, print_function
 
+import contextlib
 import os
 import re
 from itertools import chain
 
 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 inspire_dojson.cds.model import cds2hep_marc
+from inspire_dojson.utils import force_single_element, quote_url
 
 CATEGORIES = {
     'Accelerators and Storage Rings': 'Accelerators',
@@ -55,7 +53,6 @@
     'Engineering': 'Instrumentation',
     'General Relativity and Cosmology': 'Gravitation and Cosmology',
     'General Theoretical Physics': 'General Physics',
-    'General Theoretical Physics': 'General Physics',
     'Information Transfer and Management': 'Other',
     'Mathematical Physics and Mathematics': 'Math and Math Physics',
     'Nuclear Physics - Experiment': 'Experiment-Nucl',
@@ -198,10 +195,8 @@ def languages(self, key, value):
         try:
             languages.append({'a': pycountry.languages.get(alpha_3=alpha_3).name})
         except KeyError:
-            try:
+            with contextlib.suppress(KeyError):
                 languages.append({'a': pycountry.languages.get(bibliographic=alpha_3).name})
-            except KeyError:
-                pass
 
     return languages
 

inspire_dojson/common/__init__.py

@@ -24,4 +24,4 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import rules  # noqa: F401
+from inspire_dojson.common import rules  # noqa: F401

inspire_dojson/common/rules.py

@@ -27,29 +27,22 @@
 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
-from ..experiments.model import experiments
-from ..hep.model import hep, hep2marc
-from ..hepnames.model import hepnames, hepnames2marc
-from ..institutions.model import institutions
-from ..journals.model import journals
-from ..utils import (
-    force_single_element,
-    get_recid_from_ref,
-    get_record_ref,
-)
-
+from inspire_dojson.conferences.model import conferences
+from inspire_dojson.data.model import data
+from inspire_dojson.experiments.model import experiments
+from inspire_dojson.hep.model import hep, hep2marc
+from inspire_dojson.hepnames.model import hepnames, hepnames2marc
+from inspire_dojson.institutions.model import institutions
+from inspire_dojson.journals.model import journals
+from inspire_dojson.utils import force_single_element, get_recid_from_ref, get_record_ref
 
 IS_INTERNAL_UID = re.compile(r'^(inspire:uid:)?\d{5}$')
 IS_ORCID = re.compile(r'^(orcid:)?\d{4}-\d{4}-\d{4}-\d{3}[0-9X]$')
@@ -646,15 +639,9 @@ def _get_datetime(value):
     a_values = force_list(value.get('a'))
     for a_value in a_values:
         if IS_INTERNAL_UID.match(a_value):
-            if a_value.startswith('inspire:uid:'):
-                internal_uid = int(a_value[12:])
-            else:
-                internal_uid = int(a_value)
+            internal_uid = int(a_value[12:]) if a_value.startswith('inspire:uid:') else int(a_value)
         elif IS_ORCID.match(a_value):
-            if a_value.startswith('orcid:'):
-                orcid = a_value[6:]
-            else:
-                orcid = a_value
+            orcid = a_value[6:] if a_value.startswith('orcid:') else a_value
         else:
             source = a_value
 

inspire_dojson/conferences/__init__.py

@@ -24,5 +24,5 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import rules  # noqa: F401
-from .model import conferences  # noqa: F401
+from inspire_dojson.conferences import rules  # noqa: F401
+from inspire_dojson.conferences.model import conferences  # noqa: F401

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 inspire_dojson.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 inspire_dojson.conferences.model import conferences
+from inspire_dojson.utils import force_single_element
+from inspire_dojson.utils.geo import parse_conference_address
 
 
 def _trim_date(date):

inspire_dojson/data/__init__.py

@@ -24,5 +24,5 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import rules  # noqa: F401
-from .model import data  # noqa: F401
+from inspire_dojson.data import rules  # noqa: F401
+from inspire_dojson.data.model import data  # noqa: F401

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 inspire_dojson.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 inspire_dojson.data.model import data
+from inspire_dojson.utils import force_single_element, get_record_ref
 
 
 @data.over('dois', '^0247.')

inspire_dojson/experiments/__init__.py

@@ -24,5 +24,5 @@
 
 from __future__ import absolute_import, division, print_function
 
-from . import rules  # noqa: F401
-from .model import experiments  # noqa: F401
+from inspire_dojson.experiments import rules  # noqa: F401
+from inspire_dojson.experiments.model import experiments  # noqa: F401

inspire_dojson/experiments/model.py

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

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 inspire_dojson.experiments.model import experiments
+from inspire_dojson.utils import force_single_element, get_record_ref
 
 EXPERIMENT_CATEGORIES_MAP = \
     {'1': 'Collider Experiments',

inspire_dojson/hep/__init__.py

@@ -24,7 +24,8 @@
 
 from __future__ import absolute_import, division, print_function
 
-from .rules import (  # noqa: F401
+from inspire_dojson.hep.model import hep, hep2marc  # noqa: F401
+from inspire_dojson.hep.rules import (  # noqa: F401
     bd0xx,
     bd1xx,
     bd2xx,
@@ -36,4 +37,3 @@
     bd9xx,
     bdFFT,
 )
-from .model import hep, hep2marc  # noqa: F401

inspire_dojson/hep/model.py

@@ -27,16 +27,15 @@
 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
 
-from ..model import FilterOverdo, add_schema, clean_marc, clean_record
+from inspire_dojson.model import FilterOverdo, add_schema, clean_marc, clean_record
 
 
 def add_arxiv_categories(record, blob):

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 inspire_dojson.hep.model import hep, hep2marc
+from inspire_dojson.utils import force_single_element, normalize_isbn
 
 RE_LANGUAGE = re.compile(r'\/| or | and |,|=|\s+')
 

inspire_dojson/hep/rules/bd1xx.py

@@ -27,16 +27,11 @@
 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 inspire_dojson.hep.model import hep, hep2marc
+from inspire_dojson.utils import force_single_element, get_record_ref
 
 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 inspire_dojson.hep.model import hep, hep2marc
+from inspire_dojson.utils import normalize_date_aggressively
 
 
 @hep.over('titles', '^(210|245|246|247)..')