style: update ruff version + add more rules

.pre-commit-config.yaml

@@ -10,7 +10,7 @@ repos:
       - id: detect-aws-credentials
         args: [ --allow-missing-credentials ]
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.2.0  # ruff version
+    rev: v0.5.0  # ruff version
     hooks:
       - id: ruff-format
       - id: ruff

Pipfile

@@ -27,6 +27,6 @@ pytest-cov = "*"
 pytest-asyncio = "*"
 mock = "*"
 pre-commit = ">=3.7.1"
-ruff = "==0.2.0"
+ruff = "==0.5.0"
 ipykernel = "*"
 jupyterlab = "*"

pyproject.toml

@@ -43,7 +43,7 @@ dynamic = ["version"]
 
 [project.optional-dependencies]
 tests = ["pytest", "pytest-cov", "mock", "pytest-asyncio"]
-dev = ["pre-commit>=3.7.1", "ruff==0.2.0"]
+dev = ["pre-commit>=3.7.1", "ruff==0.5.0"]
 notebooks = ["ipykernel", "jupyterlab"]
 docs = [
     "sphinx==6.1.3",
@@ -104,16 +104,22 @@ select = [
     "DTZ",  # https://docs.astral.sh/ruff/rules/#flake8-datetimez-dtz
     "T10",  # https://docs.astral.sh/ruff/rules/#flake8-datetimez-dtz
     "EM",  # https://docs.astral.sh/ruff/rules/#flake8-errmsg-em
+    "LOG",  # https://docs.astral.sh/ruff/rules/#flake8-logging-log
     "G",  # https://docs.astral.sh/ruff/rules/#flake8-logging-format-g
+    "INP",  # https://docs.astral.sh/ruff/rules/#flake8-no-pep420-inp
     "PIE",  # https://docs.astral.sh/ruff/rules/#flake8-pie-pie
     "T20",  # https://docs.astral.sh/ruff/rules/#flake8-print-t20
     "PT",  # https://docs.astral.sh/ruff/rules/#flake8-pytest-style-pt
     "Q",  # https://docs.astral.sh/ruff/rules/#flake8-quotes-q
     "RSE",  # https://docs.astral.sh/ruff/rules/#flake8-raise-rse
     "RET",  # https://docs.astral.sh/ruff/rules/#flake8-return-ret
+    "SLF",  # https://docs.astral.sh/ruff/rules/#flake8-self-slf
     "SIM",  # https://docs.astral.sh/ruff/rules/#flake8-simplify-sim
+    "ARG",  # https://docs.astral.sh/ruff/rules/#flake8-unused-arguments-arg
     "PTH",  # https://docs.astral.sh/ruff/rules/#flake8-use-pathlib-pth
     "PGH",  # https://docs.astral.sh/ruff/rules/#pygrep-hooks-pgh
+    "PERF",  # https://docs.astral.sh/ruff/rules/#perflint-perf
+    "FURB",  # https://docs.astral.sh/ruff/rules/#refurb-furb
     "RUF",  # https://docs.astral.sh/ruff/rules/#ruff-specific-rules-ruf
 ]
 fixable = [
@@ -124,15 +130,19 @@ fixable = [
     "ANN",
     "B",
     "C4",
+    "LOG",
     "G",
     "PIE",
     "PT",
     "RSE",
     "SIM",
+    "PERF",
+    "FURB",
     "RUF"
 ]
-# ANN101 - missing-type-self
 # ANN003 - missing-type-kwargs
+# ANN101 - missing-type-self
+# ANN102 - missing-type-cls
 # D203 - one-blank-line-before-class
 # D205 - blank-line-after-summary
 # D206 - indent-with-spaces*
@@ -148,7 +158,7 @@ fixable = [
 # S321 - suspicious-ftp-lib-usage
 # *ignored for compatibility with formatter
 ignore = [
-    "ANN101", "ANN003",
+    "ANN003", "ANN101", "ANN102",
     "D203", "D205", "D206", "D213", "D300", "D400", "D415",
     "E111", "E114", "E117", "E501",
     "W191",
@@ -160,7 +170,9 @@ ignore = [
 # ANN2 - missing-return-type
 # ANN102 - missing-type-cls
 # S101 - assert
-# B011 - assert-false
-# N815 -
-"tests/*" = ["ANN001", "ANN2", "ANN102", "S101", "B011"]
+# INP001 - implicit-namespace-package
+# ARG001 - unused-function-argument
+# SLF001 - private-member-acces
+# N815 - mixed-case-variable-in-class-scope
+"tests/*" = ["ANN001", "ANN2", "ANN102", "S101", "INP001", "SLF001", "ARG001"]
 "src/metakb/schemas/*" = ["ANN102", "N815"]

src/metakb/__init__.py

@@ -1,4 +1,5 @@
 """The MetaKB package."""
+
 from importlib.metadata import PackageNotFoundError, version
 from pathlib import Path
 

src/metakb/cli.py

@@ -1,6 +1,7 @@
 """Provide CLI utility for performing data collection, transformation, and upload
 to graph datastore.
 """
+
 import datetime
 import logging
 import re

src/metakb/database.py

@@ -1,4 +1,5 @@
 """Acquire connection to Neo4j graph database."""
+
 import ast
 import logging
 from os import environ

src/metakb/harvesters/base.py

@@ -1,4 +1,5 @@
 """A module for the Harvester base class"""
+
 import datetime
 import json
 import logging

src/metakb/harvesters/civic.py

@@ -1,4 +1,5 @@
 """A module for the CIViC harvester."""
+
 import logging
 from typing import Any
 

src/metakb/harvesters/moa.py

@@ -1,4 +1,5 @@
 """A module for the Molecular Oncology Almanac harvester"""
+
 import logging
 
 import requests

src/metakb/load_data.py

@@ -1,12 +1,12 @@
 """Provide methods for loading data into the database."""
+
 import json
 import logging
 from pathlib import Path
 
 from neo4j import Driver, ManagedTransaction
 
 from metakb.database import get_driver
-from metakb.schemas.app import SourceName
 
 _logger = logging.getLogger(__name__)
 
@@ -456,13 +456,12 @@ def _add_study(tx: ManagedTransaction, study_in: dict) -> None:
     tx.run(query, **study)
 
 
-def add_transformed_data(driver: Driver, data: dict, src_name: SourceName) -> None:
+def add_transformed_data(driver: Driver, data: dict) -> None:
     """Add set of data formatted per Common Data Model to DB.
 
     :param data: contains key/value pairs for data objects to add to DB, including
         studies, variation, therapeutic procedures, conditions, genes, methods,
         documents, etc.
-    :param src_name: Name of source for `data`
     """
     # Used to keep track of IDs that are in studies. This is used to prevent adding
     # nodes that aren't associated to studies
@@ -508,5 +507,4 @@ def load_from_json(src_transformed_cdm: Path, driver: Driver | None = None) -> N
         driver = get_driver()
     with src_transformed_cdm.open() as f:
         items = json.load(f)
-        src_name = SourceName(str(src_transformed_cdm).split("/")[-1].split("_cdm")[0])
-        add_transformed_data(driver, items, src_name)
+        add_transformed_data(driver, items)

src/metakb/log_handle.py

@@ -5,6 +5,7 @@
 a way to manage logs in our own production environments, so the entry points that we
 define in the library make use of methods here to set some of our preferred baselines.
 """
+
 import logging
 import os
 

src/metakb/main.py

@@ -1,4 +1,5 @@
 """Main application for FastAPI."""
+
 from collections.abc import AsyncGenerator
 from contextlib import asynccontextmanager
 from typing import Annotated
@@ -14,7 +15,7 @@
 
 
 @asynccontextmanager
-async def lifespan(app: FastAPI) -> AsyncGenerator:
+async def lifespan(app: FastAPI) -> AsyncGenerator:  # noqa: ARG001
     """Configure FastAPI instance lifespan.
 
     :param app: FastAPI app instance

src/metakb/normalizers.py

@@ -1,4 +1,5 @@
 """Handle construction of and relay requests to VICC normalizer services."""
+
 import logging
 import os
 from collections.abc import Iterable

src/metakb/query.py

@@ -1,4 +1,5 @@
 """Provide class/methods/schemas for issuing queries against the database."""
+
 import json
 import logging
 from copy import copy
@@ -521,10 +522,10 @@ def _get_variations(self, cv_id: str, relation: VariationRelation) -> list[dict]
                     v_params[variation_k] = json.loads(variation_v)
                 elif variation_k.startswith("expression_hgvs_"):
                     syntax = variation_k.split("expression_")[-1].replace("_", ".")
-                    for hgvs_expr in variation_v:
-                        expressions.append(
-                            models.Expression(syntax=syntax, value=hgvs_expr)
-                        )
+                    expressions.extend(
+                        models.Expression(syntax=syntax, value=hgvs_expr)
+                        for hgvs_expr in variation_v
+                    )
 
             v_params["expressions"] = expressions or None
             loc_params = r_params["loc"]

src/metakb/schemas/annotation.py

@@ -1,4 +1,5 @@
 """Module containing GK pilot annotation definitions"""
+
 import datetime
 from enum import Enum
 from typing import Literal
@@ -23,7 +24,7 @@ class Direction(str, Enum):
     NONE = "none"
 
 
-class Document(core_models._MappableEntity):
+class Document(core_models._MappableEntity):  # noqa: SLF001
     """a representation of a physical or digital document"""
 
     type: Literal["Document"] = "Document"
@@ -41,7 +42,7 @@ class Document(core_models._MappableEntity):
     )
 
 
-class Method(core_models._Entity):
+class Method(core_models._Entity):  # noqa: SLF001
     """A set of instructions that specify how to achieve some objective (e.g.
     experimental protocols, curation guidelines, rule sets, etc.)
     """
@@ -56,7 +57,7 @@ class Method(core_models._Entity):
     )
 
 
-class Agent(core_models._Entity):
+class Agent(core_models._Entity):  # noqa: SLF001
     """An autonomous actor (person, organization, or computational agent) that bears
     some form of responsibility for an activity taking place, for the existence of an
     entity, or for another agent's activity.
@@ -67,7 +68,7 @@ class Agent(core_models._Entity):
     subtype: AgentSubtype | None = None
 
 
-class Contribution(core_models._Entity):
+class Contribution(core_models._Entity):  # noqa: SLF001
     """The sum of all actions taken by a single agent in contributing to the creation,
     modification, assessment, or deprecation of a particular entity (e.g. a Statement,
     EvidenceLine, DataItem, Publication, etc.)
@@ -98,7 +99,7 @@ def date_format(cls, v: str | None) -> str | None:
         return v
 
 
-class _InformationEntity(core_models._Entity):
+class _InformationEntity(core_models._Entity):  # noqa: SLF001
     """InformationEntities are abstract (non-physical) entities that are about something
     (i.e. they carry information about things in the real world).
     """

src/metakb/schemas/api.py

@@ -1,4 +1,5 @@
 """Create schemas for API"""
+
 from typing import Literal
 
 from pydantic import BaseModel, ConfigDict, StrictStr
@@ -12,9 +13,9 @@ class ServiceMeta(BaseModel):
 
     name: Literal["metakb"] = "metakb"
     version: StrictStr = __version__
-    url: Literal[
+    url: Literal["https://github.com/cancervariants/metakb"] = (
         "https://github.com/cancervariants/metakb"
-    ] = "https://github.com/cancervariants/metakb"
+    )
 
     model_config = ConfigDict(
         json_schema_extra={

src/metakb/schemas/app.py

@@ -1,4 +1,5 @@
 """Module containing app schemas and enums"""
+
 from enum import Enum
 
 

src/metakb/schemas/categorical_variation.py

@@ -3,6 +3,7 @@
 See the `CatVar page <https://www.ga4gh.org/product/categorical-variation-catvar/>`_ on
 the GA4GH website for more information.
 """
+
 from enum import Enum
 from typing import Literal
 
@@ -27,7 +28,7 @@ class LocationMatchCharacteristic(str, Enum):
     SUPERINTERVAL = "superinterval"
 
 
-class _CategoricalVariationBase(core_models._DomainEntity):
+class _CategoricalVariationBase(core_models._DomainEntity):  # noqa: SLF001
     """Base class for Categorical Variation"""
 
     members: list[models.Variation | core_models.IRI] | None = Field(
@@ -130,7 +131,12 @@ class CategoricalVariation(RootModel):
     individual contextual variation instances may be members of the domain.
     """
 
-    root: CanonicalAllele | CategoricalCnv | DescribedVariation | ProteinSequenceConsequence = Field(
+    root: (
+        CanonicalAllele
+        | CategoricalCnv
+        | DescribedVariation
+        | ProteinSequenceConsequence
+    ) = Field(
         ...,
         json_schema_extra={
             "description": "A representation of a categorically-defined domain for variation, in which individual contextual variation instances may be members of the domain.",

src/metakb/schemas/variation_statement.py

@@ -1,4 +1,5 @@
 """Module containing variant statement definitions"""
+
 from enum import Enum
 from typing import Literal
 

src/metakb/transform/__init__.py

@@ -1,3 +1,4 @@
 """Transformations for sources."""
+
 from .civic import CivicTransform  # noqa: F401
 from .moa import MoaTransform  # noqa: F401

src/metakb/transform/base.py

@@ -1,4 +1,5 @@
 """A module for the Transform base class."""
+
 import datetime
 import json
 import logging

src/metakb/transform/civic.py

@@ -1,4 +1,5 @@
 """A module for to transform CIViC."""
+
 import logging
 import re
 from enum import Enum
@@ -558,15 +559,14 @@ async def _add_variations(self, variants: list[dict]) -> None:
             members = await self._get_variation_members(variant)
 
             # Get variant types
-            variant_types_value = []
-            for vt in variant["variant_types"]:
-                variant_types_value.append(
-                    Coding(
-                        code=vt["so_id"],
-                        system=f"{vt['url'].rsplit('/', 1)[0]}/",
-                        label="_".join(vt["name"].lower().split()),
-                    )
+            variant_types_value = [
+                Coding(
+                    code=vt["so_id"],
+                    system=f"{vt['url'].rsplit('/', 1)[0]}/",
+                    label="_".join(vt["name"].lower().split()),
                 )
+                for vt in variant["variant_types"]
+            ]
 
             # Get mappings
             mappings = [
@@ -590,16 +590,16 @@ async def _add_variations(self, variants: list[dict]) -> None:
                     )
                 )
 
-            for ce in variant["clinvar_entries"]:
-                mappings.append(
-                    Mapping(
-                        coding=Coding(
-                            code=ce,
-                            system="https://www.ncbi.nlm.nih.gov/clinvar/variation/",
-                        ),
-                        relation=Relation.RELATED_MATCH,
-                    )
+            mappings.extend(
+                Mapping(
+                    coding=Coding(
+                        code=ce,
+                        system="https://www.ncbi.nlm.nih.gov/clinvar/variation/",
+                    ),
+                    relation=Relation.RELATED_MATCH,
                 )
+                for ce in variant["clinvar_entries"]
+            )
 
             aliases = []
             for a in variant["variant_aliases"]: