Release v09

.mise.toml

@@ -1,4 +1,4 @@
 [tools]
-python="3.11"
+python="3.12"
 poetry="2.3.3"
 java="liberica-1.8.0"

pyproject.toml

@@ -30,21 +30,21 @@ packages = [
 ]
 
 [tool.poetry.dependencies]
-python = ">=3.10,<3.12"
+python = ">=3.10,<3.13"  # breaking changes beyond 3.12
 boto3 = ">=1.34.162,<1.36"  # breaking change beyond 1.36
 botocore = ">=1.34.162,<1.36"  # breaking change beyond 1.36
-delta-spark = "2.4.*"
-duckdb = "1.1.*"  # breaking changes beyond 1.1
-Jinja2 = "3.1.*"
-lxml = "^4.9.1"
+delta-spark = "3.2.0"
+duckdb = "1.1.3"  # breaking changes beyond 1.1
+Jinja2 = "3.1.6"
+lxml = "4.9.4"
 numpy = "1.26.4"
-openpyxl = "^3.1"
-pandas = "^2.2.2"
-polars = "0.20.*"
-pyarrow = "^17.0.0"
-pydantic = "1.10.16"
-pyspark = "3.4.*"
-typing_extensions = "^4.6.2"
+openpyxl = "3.1.5"
+pandas = "2.3.3"
+polars = "0.20.31"
+pyarrow = "17.0.0"
+pydantic = "2.13.4"
+pyspark = "3.5.2"
+typing_extensions = "4.15.0"
 
 [tool.poetry.group.dev]
 optional = true
@@ -58,30 +58,30 @@ commitizen = "4.9.1"
 pre-commit = "4.3.0"
 charset-normalizer = "3.4.6"
 python-discovery = "1.2.0"
-requests = "2.33.0"
 
 [tool.poetry.group.test]
 optional = true
 
 [tool.poetry.group.test.dependencies]
-faker = "18.11.1"
 behave = "1.3.3"
 coverage = "7.11.0"
-moto = {extras = ["s3"], version = "4.0.13"}
+moto = {extras = ["s3"], version = "4.2.14"}
+requests = "2.33.0"  # dependency of `moto`
 Werkzeug = "3.1.6"
 pytest = "8.4.2"
 pytest-lazy-fixtures = "1.4.0"  # switched from https://github.com/TvoroG/pytest-lazy-fixture as it's no longer supported
-xlsx2csv = "0.8.2"
+xlsx2csv = "0.8.4"  # polars requirement
 
 [tool.poetry.group.lint]
 optional = true
 
 [tool.poetry.group.lint.dependencies]
 black = "24.3.0"
-astroid = "2.14.2"
-isort = "5.11.5"
-pylint = "2.16.4"
-mypy = "0.991"
+astroid = "3.3.9"
+isort = "5.13.2"
+pylint = "3.3.9"
+mypy = "1.20.2"
+librt = "0.11.0"  # mypy dependency
 boto3-stubs = {extras = ["essential"], version = "1.26.72"}
 botocore-stubs = "1.29.72"
 pandas-stubs = "1.2.0.62"
@@ -100,7 +100,7 @@ optional = true
 
 [tool.poetry.group.docs.dependencies]
 click = "8.2.1"
-mkdocs = "^1.6.1"
+mkdocs = "1.6.1"
 mkdocstrings = { version = "1.0.3", extras = ["python"] }
 griffelib = "2.0.1"
 pymdown-extensions = "10.21.2"
@@ -142,10 +142,6 @@ source_pkgs = [
 [tool.coverage.report]
 show_missing = true
 
-[tool.pylint]
-init-hook = "import sys; sys.path.append('./pylint_checkers')"
-load-plugins = "check_typing_imports"
-
 [tool.pylint.main]
 extension-pkg-allow-list = ["pyspark", "lxml", "pydantic"]
 fail-under = 10.0
@@ -194,7 +190,7 @@ max-statements = 50
 min-public-methods = 2
 
 [tool.pylint.exceptions]
-overgeneral-exceptions = ["BaseException", "Exception"]
+overgeneral-exceptions = ["builtins.BaseException", "builtins.Exception"]
 
 [tool.pylint.format]
 ignore-long-lines = "^\\s*(# )?<?https?://\\S+>?$"
@@ -229,6 +225,8 @@ disable = [
     "use-symbolic-message-instead",
     "logging-fstring-interpolation",
     "fixme",
+    "too-many-positional-arguments",
+    "too-many-arguments",
 ]
 enable = ["c-extension-no-member"]
 

src/dve/core_engine/backends/base/auditing.py

@@ -14,7 +14,7 @@
 from types import TracebackType
 from typing import Any, ClassVar, Generic, Optional, TypeVar, Union
 
-from pydantic import ValidationError, validate_arguments
+from pydantic import ValidationError, validate_call
 from typing_extensions import Literal, get_origin
 
 from dve.core_engine.models import (
@@ -98,8 +98,8 @@ def __init__(self, name: str, record_type: type[AuditRecord]):
     def schema(self) -> dict[str, type]:
         """Determine python schema of auditor"""
         return {
-            fld: str if get_origin(mdl.type_) == Literal else mdl.type_
-            for fld, mdl in self._record_type.__fields__.items()
+            fld: str if get_origin(mdl.annotation) == Literal else mdl.annotation  # type: ignore
+            for fld, mdl in self._record_type.model_fields.items()
         }
 
     @staticmethod
@@ -195,7 +195,7 @@ def conv_to_iterable(recs: Union[AuditorType, AuditReturnType]) -> Iterable[dict
         """Convert AuditReturnType to iterable of dictionaries"""
         raise NotImplementedError()
 
-    @validate_arguments
+    @validate_call
     def add_processing_records(self, processing_records: list[ProcessingStatusRecord]):
         """Add an entry to the processing_status auditor."""
         if self.pool:
@@ -207,7 +207,7 @@ def add_processing_records(self, processing_records: list[ProcessingStatusRecord
             records=[dict(rec) for rec in processing_records]
         )
 
-    @validate_arguments
+    @validate_call
     def add_submission_statistics_records(self, sub_stats: list[SubmissionStatisticsRecord]):
         """Add an entry to the submission statistics auditor."""
         if self.pool:
@@ -217,7 +217,7 @@ def add_submission_statistics_records(self, sub_stats: list[SubmissionStatistics
             )
         return self._submission_statistics.add_records(records=[dict(rec) for rec in sub_stats])
 
-    @validate_arguments
+    @validate_call
     def add_transfer_records(self, transfer_records: list[TransferRecord]):
         """Add an entry to the transfers auditor"""
         if self.pool:
@@ -226,7 +226,7 @@ def add_transfer_records(self, transfer_records: list[TransferRecord]):
             )
         return self._transfers.add_records(records=[dict(rec) for rec in transfer_records])
 
-    @validate_arguments
+    @validate_call
     def add_new_submissions(
         self,
         submissions: list[SubmissionMetadata],
@@ -249,7 +249,7 @@ def add_new_submissions(
                     processing_status="received",
                     job_run_id=job_run_id,
                     **ts_info,
-                ).dict(),
+                ).model_dump(),
             }
             processing_status_recs.append(processing_rec)
             if sub_info:

src/dve/core_engine/backends/implementations/duckdb/auditing.py

@@ -14,10 +14,10 @@
     OrderCriteria,
 )
 from dve.core_engine.backends.implementations.duckdb.duckdb_helpers import (
-    PYTHON_TYPE_TO_DUCKDB_TYPE,
+    get_duckdb_type_from_annotation,
     table_exists,
 )
-from dve.core_engine.backends.utilities import PYTHON_TYPE_TO_POLARS_TYPE
+from dve.core_engine.backends.utilities import get_polars_type_from_annotation
 from dve.core_engine.models import (
     AuditRecord,
     ProcessingStatusRecord,
@@ -62,18 +62,18 @@ def ddb_create_table_sql(self) -> str:
         """Generate create table sql script for auditor"""
         _sql_expression = f"CREATE TABLE {self._name} ("
         _sql_expression += ", ".join(
-            [f"{fld} {PYTHON_TYPE_TO_DUCKDB_TYPE.get(dtype)}" for fld, dtype in self.schema.items()]
+            [
+                f"{fld} {get_duckdb_type_from_annotation(dtype)}"
+                for fld, dtype in self.schema.items()
+            ]
         )
         _sql_expression += ")"
         return _sql_expression
 
     @property
     def polars_schema(self) -> dict[str, PolarsType]:
         """Get polars dataframe schema for auditor"""
-        return {
-            fld: PYTHON_TYPE_TO_POLARS_TYPE.get(dtype, pl.Utf8)  # type: ignore
-            for fld, dtype in self.schema.items()
-        }
+        return {fld: get_polars_type_from_annotation(dtype) for fld, dtype in self.schema.items()}
 
     def get_relation(self) -> DuckDBPyRelation:
         """Get a relation to interact with the auditor duckdb table"""
@@ -106,7 +106,7 @@ def conv_to_entity(self, recs: list[AuditRecord]) -> DuckDBPyRelation:
         """Convert a list of audit records to a relation"""
         # pylint: disable=W0612
         rec_df = pl.DataFrame(  # type: ignore
-            [rec.dict() for rec in recs],
+            [rec.model_dump() for rec in recs],
             schema=self.polars_schema,
         )
         return self._connection.sql("select * from rec_df")

src/dve/core_engine/backends/implementations/duckdb/contract.py

@@ -14,7 +14,6 @@
 from duckdb.typing import DuckDBPyType
 from polars.datatypes.classes import DataTypeClass as PolarsType
 from pydantic import BaseModel
-from pydantic.fields import ModelField
 
 import dve.parser.file_handling as fh
 from dve.common.error_utils import (
@@ -96,8 +95,8 @@ def create_entity_from_py_iterator(  # pylint: disable=unused-argument
     ) -> DuckDBPyRelation:
         """Create DuckDB Relation from iterator of records"""
         polars_schema: dict[str, PolarsType] = {
-            fld.name: get_polars_type_from_annotation(fld.type_)
-            for fld in stringify_model(schema).__fields__.values()
+            name: get_polars_type_from_annotation(fld.annotation)
+            for name, fld in stringify_model(schema).model_fields.items()
         }
         _lazy_df = pl.LazyFrame(records, polars_schema)  # type: ignore # pylint: disable=unused-variable
         return self._connection.sql("select * from _lazy_df")
@@ -130,17 +129,15 @@ def apply_data_contract(
         ) as msg_writer:
             for entity_name, relation in entities.items():
                 # get dtypes for all fields -> python data types or use with relation
-                entity_fields: dict[str, ModelField] = contract_metadata.schemas[
-                    entity_name
-                ].__fields__
+                entity_fields = contract_metadata.schemas[entity_name].model_fields
                 ddb_schema: dict[str, DuckDBPyType] = {
-                    fld.name: get_duckdb_type_from_annotation(fld.annotation)
-                    for fld in entity_fields.values()
+                    name: get_duckdb_type_from_annotation(fld.annotation)
+                    for name, fld in entity_fields.items()
                 }
                 ddb_schema[RECORD_INDEX_COLUMN_NAME] = get_duckdb_type_from_annotation(int)
                 polars_schema: dict[str, PolarsType] = {
-                    fld.name: get_polars_type_from_annotation(fld.annotation)
-                    for fld in entity_fields.values()
+                    name: get_polars_type_from_annotation(fld.annotation)
+                    for name, fld in entity_fields.items()
                 }
                 polars_schema[RECORD_INDEX_COLUMN_NAME] = get_polars_type_from_annotation(int)
                 if relation_is_empty(relation):