style: add new linting checks

.github/workflows/ci.yaml

@@ -0,0 +1,28 @@
+name: CI
+
+on:
+  push:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+
+jobs:
+  style:
+    name: Check style
+    runs-on: 'ubuntu-latest'
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: 3.8
+
+      - name: Install dependencies
+        run: python3 -m pip install ".[dev]"
+
+      - name: Ruff checks
+        run: python3 -m ruff check .
+
+      - name: isort checks
+        run: python3 -m isort --check .

.pre-commit-config.yaml

@@ -0,0 +1,10 @@
+repos:
+- repo: https://github.com/charliermarsh/ruff-pre-commit
+  rev: v0.0.244
+  hooks:
+    - id: ruff
+- repo: https://github.com/pycqa/isort
+  rev: 5.6.4
+  hooks:
+    - id: isort
+      args: ["--check"]

.vscode/launch.json

@@ -18,4 +18,4 @@
             "module": "anyvar.restapi"
         }
     ]
-}
+}

.vscode/settings.json

@@ -1,3 +1,3 @@
 {
     "python.formatting.provider": "yapf"
-}
+}

bin/cv-load-timing

@@ -4,20 +4,19 @@ import logging
 import os
 import sys
 
-from biocommons.seqrepo import SeqRepo
-from bioutils.exceptions import BioutilsError
 import coloredlogs
 import redis
+import tqdm
+from biocommons.seqrepo import SeqRepo
+from bioutils.exceptions import BioutilsError
 from ga4gh.vrs import models
 from ga4gh.vrs.dataproxy import SeqRepoDataProxy
 from hgvs.exceptions import HGVSParseError
-import tqdm
 
 from anyvar import AnyVar
-from anyvar.translator import Translator
 from anyvar.extras.clinvarparser import ClinvarParser
 from anyvar.storage.redisobjectstore import RedisObjectStore
-
+from anyvar.translator import Translator
 
 _logger = logging.getLogger(__name__)
 
@@ -30,7 +29,7 @@ def firstn(gen, n=None):
             yield e
             if i == n:
                 break
-        
+
 
 
 if __name__ == "__main__":
@@ -55,17 +54,22 @@ if __name__ == "__main__":
     n_cvrec = 0
     n_exc = 0
     for cvrec in pbar:
-        #_logger.info(f"{cvrec.accession}; {cvrec.record_type}; {len(cvrec.hgvs_expressions)} hgvs expressions")
+        #_logger.info(
+        #     f"{cvrec.accession}; {cvrec.record_type}; {len(cvrec.hgvs_expressions)} "
+        #     "hgvs expressions"
+        # )
         n_cvrec += 1
         n_hgvs += len(cvrec.hgvs_expressions)
         if n_cvrec % 10 == 0:
             pbar.set_description(f"{n_hgvs} hgvs expressions")
         for he in cvrec.hgvs_expressions:
             try:
                 v = tlr.from_hgvs(he)
-            except (ValueError, HGVSParseError, AttributeError, KeyError, BioutilsError):
+            except (
+                ValueError, HGVSParseError, AttributeError, KeyError, BioutilsError
+            ):
                 v = models.Text(definition=he)
             _id = av.put_object(v)
-            #_logger.info(f"stored {_id} for {he}")
+            _logger.info(f"stored {_id} for {he}")
 
     _logger.info(f"{n_hgvs} hgvs expressions in {n_cvrec} clinvar records")

docker-build

@@ -10,5 +10,5 @@ tag=$(git describe --tags)
 e docker build -t $imagename:$tag "$@" .
 e docker push $imagename:$tag
 
-e docker tag $imagename:$tag $imagename:latest 
-e docker push $imagename:latest 
+e docker tag $imagename:$tag $imagename:latest
+e docker push $imagename:latest

docker-compose.yml

@@ -18,7 +18,7 @@ services:
       - seqrepo-rest-service
     #  - uta
     #environment:
-    #  # UTA_DB_URL: 
+    #  # UTA_DB_URL:
     #  # SEQREPO_REST_SERVICE_URL:
 
   redis:
@@ -28,7 +28,7 @@ services:
     volumes:
       - anyvar_redis_vol:/data
 
-  seqrepo-rest-service:    
+  seqrepo-rest-service:
     # Test: curl http://localhost:5000/seqrepo/1/sequence/refseq:NM_000551.3
     expose:
       - 5000
@@ -37,7 +37,7 @@ services:
     image: biocommons/seqrepo-rest-service:latest
     volumes:
       - seqrepo_vol:/usr/local/share/seqrepo
-    
+
   # uta:
   #   # Test:
   #   # psql -XAt postgres://anonymous@localhost/uta -c 'select count(*) from transcript'

pyproject.toml

@@ -0,0 +1,16 @@
+[tool.ruff]
+exclude = [
+    "venv",
+    "build",
+    "dist",
+    "tests"
+]
+
+line-length = 88
+
+[tool.ruff.per-file-ignores]
+"__init__.py" = ["F401", "E402"]
+"uidoc.py" = ["E501"]
+
+[tool.isort]
+group_by_package = true

setup.cfg

@@ -32,10 +32,13 @@ exclude =
 
 [options.extras_require]
 dev =
-	ipython
+	ruff
+        pre-commit
+        isort
 test =
 	pytest
         pytest-cov
+        pytest-mock
 
 
 [aliases]
@@ -47,12 +50,5 @@ universal = 1
 [build_sphinx]
 all_files  = 1
 
-# http://pep8.readthedocs.org/en/latest/intro.html#error-codes
-[flake8]
-max-line-length = 120
-exclude = tests/*
-max-complexity = 10
-ignore = E129,E221,E241,E251,E303,W291
-
 [tool:pytest]
 addopts = --cov=anyvar --cov-report term-missing

src/anyvar/anyvar.py

@@ -3,22 +3,23 @@
 
 """
 
-from collections.abc import MutableMapping
 import logging
+from collections.abc import MutableMapping
 
 from ga4gh.core import ga4gh_identify
 from ga4gh.vrs import models, vrs_deref, vrs_enref
 
 from anyvar.translate.translate import _Translator
 
-
 _logger = logging.getLogger(__name__)
 
 
 class AnyVar:
     def __init__(self, /, translator: _Translator, object_store: MutableMapping):
         if not isinstance(object_store, MutableMapping):
-            _logger.warning("AnyVar(object_store=) should be a mutable mapping; you're on your own")
+            _logger.warning(
+                "AnyVar(object_store=) should be a mutable mapping; you're on your own"
+            )
 
         self.object_store = object_store
         self.translator = translator

src/anyvar/extras/clinvarparser.py

@@ -3,13 +3,13 @@
 
 import lxml.etree as le
 
-
 _logger = logging.getLogger()
 
 
 class VariationArchive:
     def __init__(self, element):
-        assert element.tag == "VariationArchive", "Expected node type `VariationArchive`"
+        assert element.tag == "VariationArchive", \
+            "Expected node type `VariationArchive`"
         self._element = element
 
     @property
@@ -39,15 +39,15 @@ def record_type(self):
     @property
     def variation_name(self):
         return self._element.get("VariationName")
-    
+
     @property
     def variation_type(self):
         return self._element.get("VariationType")
 
     @property
     def version(self):
         return self._element.get("Version")
-    
+
     @property
     def xrefs(self):
         return [e.attrib for e in self._element.xpath(".//XRefList/XRef")]
@@ -75,4 +75,7 @@ def clinvar_open(fp):
     import sys
     fn = sys.argv[1]
     for va in clinvar_open(fn):
-        print(f"{va.acv}\t{va.variation_name}\t{va.variation_type}\t{len(va.hgvs_expressions)}\t{len(va.xrefs)}")
+        print(
+            f"{va.acv}\t{va.variation_name}\t{va.variation_type}\t"
+            f"{len(va.hgvs_expressions)}\t{len(va.xrefs)}"
+        )

src/anyvar/restapi/_data/openapi.yaml

@@ -615,4 +615,3 @@ components:
           length: 248956422
           alphabet: ACGMNRT
           added: 2016-08-27T21:17:00Z
-

src/anyvar/restapi/globals.py

@@ -8,10 +8,11 @@
 from typing import Any, Callable, Optional
 
 from flask import current_app
-from ga4gh.vrs.dataproxy import create_dataproxy, _DataProxy
-from anyvar.translate.translate import TranslatorSetupException
+from ga4gh.vrs.dataproxy import _DataProxy, create_dataproxy
 
-from anyvar.translate.variation_normalizer import VariationNormalizerRestTranslator
+from anyvar.translate.translate import TranslatorSetupException
+from anyvar.translate.variation_normalizer import \
+    VariationNormalizerRestTranslator
 
 from ..anyvar import AnyVar
 from ..storage import create_storage

src/anyvar/restapi/routes/allele.py

@@ -15,7 +15,9 @@ def put(body):
     except TranslationException:
         result["messages"].append(f"Unable to translate {defn}")
     except NotImplementedError:
-        result["messages"].append(f"Variation class for {defn} is currently unsupported.")
+        result["messages"].append(
+            f"Variation class for {defn} is currently unsupported."
+        )
     else:
         v_id = av.put_object(v)
         result["object"] = v.as_dict()

src/anyvar/restapi/routes/find_alleles.py

@@ -1,7 +1,7 @@
-from connexion import NoContent
+from anyvar.storage.postgres import PostgresObjectStore
 
 from ..globals import get_anyvar
-from anyvar.storage.postgres import PostgresObjectStore
+
 
 def get_ga4gh_alias(seqrepo_data_proxy, accession):
     md = seqrepo_data_proxy.get_metadata(accession)

src/anyvar/restapi/routes/info.py

@@ -1,8 +1,6 @@
-from connexion import NoContent
-
-import anyvar
 import ga4gh.vrs
 
+import anyvar
 
 
 def search():

src/anyvar/restapi/routes/sequence.py

@@ -1,7 +1,7 @@
 import logging
 
 from ..globals import get_dataproxy
-from .utils import get_sequence_ids, problem
+from .utils import problem
 
 _logger = logging.getLogger(__name__)
 
@@ -12,4 +12,3 @@ def get(alias, start=None, end=None):
             return problem(422, "Invalid coordinates: start > end")
     dp = get_dataproxy()
     return dp.get_sequence(alias, start, end), 200
-

src/anyvar/restapi/routes/sequence_metadata.py

@@ -1,10 +1,6 @@
 import logging
 
-from connexion import NoContent
-
 from ..globals import get_dataproxy
-from .utils import get_sequence_ids, problem
-
 
 _logger = logging.getLogger(__name__)
 

src/anyvar/restapi/routes/summary_statistics.py

@@ -1,5 +1,6 @@
 from ..globals import get_anyvar
 
+
 def get(vartype):
     av = get_anyvar()
     if vartype == "substitution":

src/anyvar/restapi/routes/text.py

@@ -1,5 +1,3 @@
-from connexion import NoContent
-
 from ..globals import get_anyvar
 
 
@@ -8,7 +6,7 @@ def put(body):
     request = body
     defn = request.pop("definition")
     v = av.create_text(defn)
-    id = av.put_object(v)
+    av.put_object(v)
     result = {
         "object": v.as_dict(),
         "messages": [],