[pre-commit.ci] pre-commit autoupdate

.pre-commit-config.yaml

@@ -4,7 +4,7 @@
 files: ^soda-[^/]+/
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.4.0
+    rev: v6.0.0
     hooks:
       - id: trailing-whitespace
       - id: check-added-large-files
@@ -17,7 +17,7 @@ repos:
       - id: detect-private-key
       - id: end-of-file-fixer
   - repo: https://github.com/PyCQA/autoflake
-    rev: v2.2.1
+    rev: v2.3.3
     hooks:
       - id: autoflake
         args: ["--in-place", "--remove-all-unused-imports"]
@@ -29,13 +29,13 @@ repos:
   #        exclude: _models?\.py$
   #        args: [--py38-plus, --keep-runtime-typing]
   - repo: https://github.com/PyCQA/isort
-    rev: 5.12.0
+    rev: 9.0.0a3
     hooks:
       - id: isort
         additional_dependencies: [toml]
         name: Sort imports using isort
-  - repo: https://github.com/psf/black
-    rev: 23.3.0
+  - repo: https://github.com/psf/black-pre-commit-mirror
+    rev: 26.3.1
     hooks:
       - id: black
         name: Run black formatter

soda-core/src/soda_core/cli/handlers/data_source.py

@@ -23,19 +23,15 @@ def handle_create_data_source(data_source_file_path: str, data_source_type: str)
     try:
         Path(dir_name).mkdir(parents=True, exist_ok=True)
         with open(data_source_file_path, "w") as text_file:
-            text_file.write(
-                dedent(
-                    """
+            text_file.write(dedent("""
                 type: postgres
                 name: postgres_ds
                 connection:
                     host: localhost
                     user: ${POSTGRES_USERNAME}
                     password: ${POSTGRES_PASSWORD}
                     database: your_postgres_db
-                """
-                ).strip()
-            )
+                """).strip())
         soda_logger.info(f"{Emoticons.WHITE_CHECK_MARK} Created data source file '{data_source_file_path}'")
         return ExitCode.OK
     except Exception as exc:

soda-core/src/soda_core/cli/handlers/soda_cloud.py

@@ -20,16 +20,12 @@ def handle_create_soda_cloud(soda_cloud_file_path: str) -> ExitCode:
     try:
         Path(dirname(soda_cloud_file_path)).mkdir(parents=True, exist_ok=True)
         with open(soda_cloud_file_path, "w") as text_file:
-            text_file.write(
-                dedent(
-                    """
+            text_file.write(dedent("""
                 soda_cloud:
                   host: cloud.soda.io
                   api_key_id: ${SODA_CLOUD_API_KEY_ID}
                   api_key_secret: ${SODA_CLOUD_API_KEY_SECRET}
-                """
-                ).strip()
-            )
+                """).strip())
         soda_logger.info(f"{Emoticons.WHITE_CHECK_MARK} Created Soda Cloud configuration file '{soda_cloud_file_path}'")
         return ExitCode.OK
     except Exception as exc:

soda-core/src/soda_core/common/soda_cloud.py

@@ -1457,12 +1457,14 @@ def _build_contract_result_json_dict(contract_verification_result: ContractVerif
             # The scan definition name is still required on result ingestion to link to the contract
             # and determine if we're dealing with a default or test contract.
             "definitionName": _build_scan_definition_name(contract_verification_result),
-            "defaultDataSource": contract_verification_result.data_source.name
-            if contract_verification_result.data_source
-            else None,
-            "defaultDataSourceProperties": {"type": contract_verification_result.data_source.type}
-            if contract_verification_result.data_source
-            else None,
+            "defaultDataSource": (
+                contract_verification_result.data_source.name if contract_verification_result.data_source else None
+            ),
+            "defaultDataSourceProperties": (
+                {"type": contract_verification_result.data_source.type}
+                if contract_verification_result.data_source
+                else None
+            ),
             # dataTimestamp can be changed by user, this is shown in Cloud as time of a scan.
             # It's the timestamp used to identify the time partition, which is the slice of data that is verified.
             "dataTimestamp": contract_verification_result.data_timestamp,
@@ -1547,9 +1549,11 @@ def _build_diagnostics_json_dict(check_result: CheckResult) -> Optional[dict]:
 
     return {
         #  TODO: this default 0 value is here only because check.diagnostics.value is a required non-nullable field in the api.
-        "value": int(check_result.threshold_value)
-        if isinstance(check_result.threshold_value, bool)
-        else (check_result.threshold_value or 0),
+        "value": (
+            int(check_result.threshold_value)
+            if isinstance(check_result.threshold_value, bool)
+            else (check_result.threshold_value or 0)
+        ),
         "fail": _build_fail_threshold(check_result),
         "v4": _build_v4_diagnostics_check_type_json_dict(check_result),
     }

soda-core/src/soda_core/common/sql_ast.py

@@ -61,10 +61,8 @@ def __post_init__(self):
         # Check that the select contains a distinct and has multiple fields -> give a warning
         if isinstance(self.fields, list) and len(self.fields) > 1:
             if any(isinstance(field, DISTINCT) for field in self.fields):
-                logger.warning(
-                    """Found DISTINCT in a SELECT statement with multiple fields.
-                               This might have unintended consequences."""
-                )
+                logger.warning("""Found DISTINCT in a SELECT statement with multiple fields.
+                               This might have unintended consequences.""")
 
 
 @dataclass

soda-core/src/soda_core/common/yaml.py

@@ -181,16 +181,13 @@ def parse(self) -> YamlObject:
             raise YamlParserException(f"YAML syntax error", str(location))
 
 
-class DataSourceYamlSource(YamlSource, file_type=FileType.DATA_SOURCE):
-    ...
+class DataSourceYamlSource(YamlSource, file_type=FileType.DATA_SOURCE): ...
 
 
-class SodaCloudYamlSource(YamlSource, file_type=FileType.SODA_CLOUD):
-    ...
+class SodaCloudYamlSource(YamlSource, file_type=FileType.SODA_CLOUD): ...
 
 
-class ContractYamlSource(YamlSource, file_type=FileType.CONTRACT):
-    ...
+class ContractYamlSource(YamlSource, file_type=FileType.CONTRACT): ...
 
 
 class YamlValue:

soda-core/src/soda_core/contracts/impl/check_types/schema_check.py

@@ -291,9 +291,9 @@ def execute(self) -> list[Measurement]:
             logger.error(msg=f"Could not execute schema query {self.sql}: {e}", exc_info=True)
             return []
         try:
-            metadata_columns: list[
-                ColumnMetadata
-            ] = self.data_source_impl.sql_dialect.build_column_metadatas_from_query_result(query_result)
+            metadata_columns: list[ColumnMetadata] = (
+                self.data_source_impl.sql_dialect.build_column_metadatas_from_query_result(query_result)
+            )
         except Exception as e:
             logger.error(f"Error building column metadata from query result: {e}")
             return []

soda-core/src/soda_core/contracts/impl/contract_yaml.py

@@ -32,8 +32,7 @@
 
 class ContractYamlExtension(Protocol):
     # Extend the contract YAML object. Can modify the state of the contract YAML.
-    def extend(self, contract_yaml: "ContractYaml") -> None:
-        ...
+    def extend(self, contract_yaml: "ContractYaml") -> None: ...
 
 
 class ContractYaml:

soda-databricks/src/soda_databricks/model/data_source/databricks_connection_properties.py

@@ -7,8 +7,7 @@
 )
 
 
-class DatabricksConnectionProperties(DataSourceConnectionProperties, ABC):
-    ...
+class DatabricksConnectionProperties(DataSourceConnectionProperties, ABC): ...
 
 
 class DatabricksSharedConnectionProperties(DatabricksConnectionProperties, ABC):

soda-postgres/src/soda_postgres/common/data_sources/postgres_data_source.py

@@ -270,8 +270,7 @@ def _build_pg_columns_metadata_query(
             [
                 COLUMN("attname", table_alias="a", field_alias="column_name"),
                 # Normalize data type into information_schema.columns style.
-                RAW_SQL(
-                    """CASE
+                RAW_SQL("""CASE
                         -- arrays
                         WHEN t.typcategory = 'A' OR t.typelem <> 0 THEN 'ARRAY'
 
@@ -294,57 +293,48 @@ def _build_pg_columns_metadata_query(
                         ELSE COALESCE(bt.typname, t.typname)
                         END
                     END AS  \"data_type\"
-                """
-                ),
+                """),
                 # Extract type parameters. All a.atttypmod are offset by 4 in Postgres
                 #  varchar/char length (NULL otherwise)
-                RAW_SQL(
-                    """CASE
+                RAW_SQL("""CASE
                         WHEN t.typname IN ('varchar','bpchar') THEN
                             CASE
                                 WHEN a.atttypmod > 4 THEN a.atttypmod - 4
                                 ELSE NULL
                             END
                         ELSE NULL
                     END AS "character_maximum_length"
-                """
-                ),
+                """),
                 # numeric precision (NULL otherwise)
-                RAW_SQL(
-                    """CASE
+                RAW_SQL("""CASE
                         WHEN t.typname = 'numeric' THEN
                             CASE
                                 WHEN a.atttypmod > 4 THEN ((a.atttypmod - 4) >> 16)
                                 ELSE NULL
                             END
                         ELSE NULL
                     END AS "numeric_precision"
-                """
-                ),
+                """),
                 # numeric scale (NULL otherwise)
-                RAW_SQL(
-                    """CASE
+                RAW_SQL("""CASE
                         WHEN t.typname = 'numeric' THEN
                             CASE
                                 WHEN a.atttypmod > 4 THEN ((a.atttypmod - 4) & 65535)
                                 ELSE NULL
                             END
                         ELSE NULL
                     END AS "numeric_scale"
-                """
-                ),
+                """),
                 # datetime precision (NULL otherwise)
-                RAW_SQL(
-                    """CASE
+                RAW_SQL("""CASE
                         WHEN t.typname IN ('time','timetz','timestamp','timestamptz') THEN
                             CASE
                                 WHEN a.atttypmod >= 0 THEN a.atttypmod
                                 ELSE NULL
                             END
                         ELSE NULL
                     END AS "datetime_precision"
-                """
-                ),
+                """),
                 COLUMN(current_database_expression, field_alias="table_catalog"),
                 COLUMN("nspname", table_alias="n", field_alias="table_schema"),
                 COLUMN("relname", table_alias="c", field_alias="table_name"),

soda-snowflake/src/soda_snowflake/common/data_sources/snowflake_data_source_connection.py

@@ -21,8 +21,7 @@
 USER_DESCRIPTION = "Username for authentication"
 
 
-class SnowflakeConnectionProperties(DataSourceConnectionProperties, ABC):
-    ...
+class SnowflakeConnectionProperties(DataSourceConnectionProperties, ABC): ...
 
 
 class SnowflakeSharedConnectionProperties(SnowflakeConnectionProperties, ABC):

soda-snowflake/tests/data_sources/test_snowflake.py

@@ -92,7 +92,7 @@
     # TestConnection(  # connect using external browser. commented out by default because CI will hang.  Uncomment to test locally
     #     test_name="external_browser_connection",
     #     connection_yaml_str=f"""
-    #             type: snowflake
+    # type: snowflake
     #             name: SNOWFLAKE_TEST
     #             connection:
     #                 account: '{SNOWFLAKE_ACCOUNT}'

soda-sqlserver/src/soda_sqlserver/common/data_sources/sqlserver_data_source.py

@@ -138,9 +138,7 @@ def create_schema_if_not_exists_sql(self, prefixes: list[str], add_semicolon: bo
                         FROM    sys.schemas
                         WHERE   name = N'{schema_name}' )
         EXEC('CREATE SCHEMA [{schema_name}]')
-        """ + (
-            ";" if add_semicolon else ""
-        )
+        """ + (";" if add_semicolon else "")
 
     def build_drop_table_sql(self, drop_table: DROP_TABLE | DROP_TABLE_IF_EXISTS, add_semicolon: bool = True) -> str:
         if_exists_sql: str = (

soda-tests/tests/unit/check_types/yaml_parsing/test_aggregate_check_yaml_parsing.py

@@ -9,7 +9,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_duplicate_check_yaml_parsing.py

@@ -8,7 +8,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract, parse_column_check
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_failed_rows_check_yaml_parsing.py

@@ -11,7 +11,6 @@
 Common fields (name, filter, threshold, metric, store_failed_rows) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_freshness_check_yaml_parsing.py

@@ -11,7 +11,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_invalid_check_yaml_parsing.py

@@ -12,7 +12,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_column_check
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_metric_check_yaml_parsing.py

@@ -11,7 +11,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_missing_check_yaml_parsing.py

@@ -10,7 +10,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_column_check
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_row_count_check_yaml_parsing.py

@@ -7,7 +7,6 @@
 Common fields (name, filter, threshold, metric, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract
 
 

soda-tests/tests/unit/check_types/yaml_parsing/test_schema_check_yaml_parsing.py

@@ -10,7 +10,6 @@
 Common fields (name, qualifier, filter, attributes) are tested in test_common_check_yaml_features.py.
 """
 
-
 from helpers.yaml_parsing_helpers import parse_check_from_contract