Merge pull request #1658 from weaviate/chore/replace-black-autoflake-with-ruff

Add ruff, remove black/autoflake from pre-commit

Author: dirkkul

.github/workflows/main.yaml

@@ -37,8 +37,10 @@ jobs:
           python-version: "3.11"
           cache: 'pip' # caching pip dependencies
       - run: pip install -r requirements-devel.txt
-      - name: "Black"
-        run: black --check weaviate test mock_tests integration
+      - name: "Ruff lint"
+        run: ruff check weaviate test mock_tests integration
+      - name: "Ruff format"
+        run: ruff format --diff weaviate test mock_tests integration
       - name: "Flake 8"
         run: flake8 weaviate test mock_tests integration
       - name: "Check release for pypi"

.pre-commit-config.yaml

@@ -7,24 +7,16 @@ repos:
       language: system
       entry: ./tools/stubs_regen.sh
 
-- repo: https://github.com/psf/black-pre-commit-mirror
-  rev: 24.10.0
+- repo: https://github.com/astral-sh/ruff-pre-commit
+  # Ruff version.
+  rev: v0.11.7
   hooks:
-    - id: black
-      language_version: python3.12
-
-- repo: https://github.com/pre-commit/pre-commit-hooks
-  rev: v4.6.0
-  hooks:
-    - id: no-commit-to-branch
-    - id: trailing-whitespace
-
-- repo: https://github.com/myint/autoflake
-  rev: v2.2.1 # autoflake v2.2.1 is the latest version that supports Python 3.12
-  hooks:
-      - id: autoflake
-        args: [--in-place, --remove-all-unused-imports, --exclude=weaviate/proto/*]
-
+    # Run the linter.
+    - id: ruff
+      args: [ weaviate, integration, test, mock_tests, journey_tests, --fix ]
+    # Run the formatter.
+    - id: ruff-format
+      args: [ weaviate, integration, test, mock_tests, journey_tests ]
 
 - repo: https://github.com/PyCQA/flake8
   rev: 7.1.0
@@ -43,6 +35,12 @@ repos:
         'pydoclint==0.6.5',
       ]
 
+- repo: https://github.com/pre-commit/pre-commit-hooks
+  rev: v4.6.0
+  hooks:
+    - id: no-commit-to-branch
+    - id: trailing-whitespace
+
 - repo: local
   hooks:
     - id: pyright

integration/conftest.py

@@ -2,13 +2,13 @@
 from typing import (
     Any,
     AsyncGenerator,
-    Optional,
-    List,
+    Dict,
     Generator,
+    List,
+    Optional,
     Protocol,
-    Type,
-    Dict,
     Tuple,
+    Type,
     Union,
 )
 
@@ -19,17 +19,17 @@
 import weaviate
 from weaviate.collections import Collection, CollectionAsync
 from weaviate.collections.classes.config import (
+    Configure,
+    DataType,
     Property,
-    _VectorizerConfigCreate,
+    _GenerativeProvider,
     _InvertedIndexConfigCreate,
+    _MultiTenancyConfigCreate,
     _ReferencePropertyBase,
-    Configure,
-    _GenerativeProvider,
     _ReplicationConfigCreate,
-    DataType,
-    _MultiTenancyConfigCreate,
-    _VectorIndexConfigCreate,
     _RerankerProvider,
+    _VectorIndexConfigCreate,
+    _VectorizerConfigCreate,
 )
 from weaviate.collections.classes.config_named_vectors import _NamedVectorConfigCreate
 from weaviate.collections.classes.types import Properties

integration/test_backup_v4.py

@@ -1,7 +1,7 @@
 import datetime
 import pathlib
 import time
-from typing import Generator, List, Union, Optional
+from typing import Generator, List, Optional, Union
 
 import pytest
 
@@ -16,8 +16,8 @@
 )
 from weaviate.collections.classes.config import DataType, Property, ReferenceProperty
 from weaviate.exceptions import (
-    UnexpectedStatusCodeException,
     BackupFailedException,
+    UnexpectedStatusCodeException,
 )
 
 pytestmark = pytest.mark.xdist_group(name="backup")

integration/test_batch_v4.py

@@ -1,15 +1,15 @@
 import concurrent.futures
 import uuid
 from dataclasses import dataclass
-from typing import Generator, List, Optional, Protocol, Tuple, Callable
+from typing import Callable, Generator, List, Optional, Protocol, Tuple
 
 import pytest
 from _pytest.fixtures import SubRequest
 
 import weaviate
-from weaviate import BatchClient, ClientBatchingContextManager
 import weaviate.classes as wvc
 from integration.conftest import _sanitize_collection_name
+from weaviate import BatchClient, ClientBatchingContextManager
 from weaviate.collections.classes.batch import Shard
 from weaviate.collections.classes.config import (
     Configure,
@@ -19,8 +19,8 @@
 )
 from weaviate.collections.classes.grpc import QueryReference
 from weaviate.collections.classes.internal import (
-    _CrossReference,
     ReferenceToMulti,
+    _CrossReference,
 )
 from weaviate.collections.classes.tenants import Tenant
 from weaviate.types import UUID, VECTORS

integration/test_client.py

@@ -4,20 +4,19 @@
 from _pytest.fixtures import SubRequest
 
 import weaviate
+import weaviate.classes as wvc
 from weaviate.collections import Collection
 from weaviate.collections.classes.config import (
     Configure,
-    _CollectionConfig,
     DataType,
     GenerativeSearches,
     Property,
     ReferenceProperty,
     Vectorizers,
+    _CollectionConfig,
 )
-from weaviate.exceptions import WeaviateClosedClientError, WeaviateStartUpError
-import weaviate.classes as wvc
-
 from weaviate.config import Timeout
+from weaviate.exceptions import WeaviateClosedClientError, WeaviateStartUpError
 
 WCS_HOST = "piblpmmdsiknacjnm1ltla.c1.europe-west3.gcp.weaviate.cloud"
 WCS_URL = f"https://{WCS_HOST}"

integration/test_client_debug.py

@@ -1,11 +1,11 @@
 import pytest
+
 from integration.conftest import (
     AsyncClientFactory,
     AsyncCollectionFactory,
     ClientFactory,
     CollectionFactory,
 )
-
 from weaviate.classes.config import DataType, Property
 from weaviate.classes.debug import DebugRESTObject
 

integration/test_collection.py

@@ -8,45 +8,45 @@
 import pytest
 
 from integration.conftest import CollectionFactory, CollectionFactoryGet, _sanitize_collection_name
-from integration.constants import WEAVIATE_LOGO_OLD_ENCODED, WEAVIATE_LOGO_NEW_ENCODED
+from integration.constants import WEAVIATE_LOGO_NEW_ENCODED, WEAVIATE_LOGO_OLD_ENCODED
 from weaviate.collections.classes.batch import ErrorObject
 from weaviate.collections.classes.config import (
     Configure,
+    ConsistencyLevel,
     DataType,
     Property,
     ReferenceProperty,
     Tokenization,
     Vectorizers,
-    ConsistencyLevel,
 )
 from weaviate.collections.classes.data import (
     DataObject,
 )
 from weaviate.collections.classes.grpc import (
-    QueryReference,
-    HybridFusion,
+    PROPERTIES,
+    PROPERTY,
+    REFERENCE,
     GroupBy,
+    HybridFusion,
     MetadataQuery,
     Move,
+    NearMediaType,
+    QueryReference,
     Sort,
     _Sorting,
-    PROPERTIES,
-    PROPERTY,
-    REFERENCE,
-    NearMediaType,
 )
 from weaviate.collections.classes.internal import (
-    _CrossReference,
     Object,
     ReferenceToMulti,
+    _CrossReference,
 )
-from weaviate.collections.classes.types import PhoneNumber, _PhoneNumber, WeaviateProperties
+from weaviate.collections.classes.types import PhoneNumber, WeaviateProperties, _PhoneNumber
 from weaviate.exceptions import (
     UnexpectedStatusCodeError,
-    WeaviateInvalidInputError,
-    WeaviateQueryError,
     WeaviateInsertInvalidPropertyError,
     WeaviateInsertManyAllFailedError,
+    WeaviateInvalidInputError,
+    WeaviateQueryError,
     WeaviateUnsupportedFeatureError,
 )
 from weaviate.types import UUID, UUIDS

integration/test_collection_aggregate.py

@@ -12,24 +12,22 @@
     AggregateDate,
     AggregateInteger,
     AggregateNumber,
-    AggregateText,
     AggregateReturn,
-    Metrics,
+    AggregateText,
     GroupByAggregate,
+    Metrics,
 )
-from weaviate.collections.classes.config import DataType, Property, ReferenceProperty, Configure
+from weaviate.collections.classes.config import Configure, DataType, Property, ReferenceProperty
 from weaviate.collections.classes.filters import Filter, _Filters
+from weaviate.collections.classes.grpc import Move
+from weaviate.collections.classes.tenants import Tenant
 from weaviate.exceptions import (
     WeaviateInvalidInputError,
     WeaviateQueryError,
     WeaviateUnsupportedFeatureError,
 )
 from weaviate.util import file_encoder_b64
 
-from weaviate.collections.classes.grpc import Move
-
-from weaviate.collections.classes.tenants import Tenant
-
 UUID1 = uuid.UUID("8ad0d33c-8db1-4437-87f3-72161ca2a51a")
 UUID2 = uuid.UUID("577887c1-4c6b-5594-aa62-f0c17883d9cf")
 
@@ -201,7 +199,7 @@ def test_over_all_with_filters_ref(collection_factory: CollectionFactory) -> Non
     assert res.properties["text"].count == 1
     assert res.properties["text"].top_occurrences[0].value == "two"
 
-    query = lambda: collection.aggregate.over_all(
+    query = lambda: collection.aggregate.over_all(  # noqa: E731
         filters=Filter.by_ref("ref").by_property("text").equal("one"),
         return_metrics=[Metrics("text").text(count=True, top_occurrences_value=True)],
     )
@@ -363,7 +361,7 @@ def test_hybrid_aggregation_group_by(
     collection.data.insert({"text": text_1})
     collection.data.insert({"text": text_2})
 
-    querier = lambda: collection.aggregate.hybrid(
+    querier = lambda: collection.aggregate.hybrid(  # noqa: E731
         "text",
         alpha=0,
         query_properties=["text"],
@@ -400,7 +398,7 @@ def test_hybrid_aggregation_group_by_with_named_vectors(
     collection.data.insert({"text": text_1})
     collection.data.insert({"text": text_2})
 
-    querier = lambda: collection.aggregate.hybrid(
+    querier = lambda: collection.aggregate.hybrid(  # noqa: E731
         "text",
         alpha=0,
         query_properties=["text"],
@@ -434,7 +432,7 @@ def test_hybrid_aggregation_group_by_with_named_vectors(
 def test_near_vector_aggregation(
     collection_factory: CollectionFactory, option: dict, expected_len: int
 ) -> None:
-    collection_maker = lambda: collection_factory(
+    collection_maker = lambda: collection_factory(  # noqa: E731
         properties=[Property(name="text", data_type=DataType.TEXT)],
         vectorizer_config=Configure.Vectorizer.text2vec_contextionary(
             vectorize_collection_name=False

integration/test_collection_batch.py

@@ -1,6 +1,6 @@
 import uuid
 from dataclasses import dataclass
-from typing import Generator, Optional, Union, Any, Protocol
+from typing import Any, Generator, Optional, Protocol, Union
 
 import pytest
 
@@ -13,9 +13,8 @@
     ReferenceProperty,
 )
 from weaviate.collections.classes.grpc import QueryReference
-from weaviate.collections.classes.internal import _CrossReference, ReferenceToMulti
+from weaviate.collections.classes.internal import ReferenceToMulti, _CrossReference
 from weaviate.collections.classes.tenants import Tenant
-
 from weaviate.types import VECTORS
 
 UUID = Union[str, uuid.UUID]