GH-62: refactor(metadata): Rename ColumnProfile to ColumnInformation

The Pydantic model `ColumnProfile` and its corresponding field `column_profile` have been renamed to `ColumnInformation` and `columns` respectively.

This change improves clarity and consistency. "Information" is a more accurate and general term for the data being stored (name, role, statistics) than "Profile". The field name `columns` is more concise and conventional.

All related functions, variables, test fixtures, and test data have been updated to reflect this renaming. No functional changes are introduced.

GH-62: refactor: Add support for text column statistics

Introduces support for calculating and storing summary statistics for columns with the `text` role.

This includes:
- Adding a `ColumnStatisticsText` model.
- Refactoring `ColumnStatistics` models to use a common `ColumnStatisticsBase` class, reducing code duplication.
- Making statistical fields (like min, max, avg) optional to correctly handle columns with only NULL values.
- Introducing a `ColumnType` enum to replace string literals for column types, improving type safety and readability.
- Updating test fixtures and helpers to align with the new data models and support the `text` role.

GH-62: ci: Optimize workflow triggers with path filtering

Adds path filters to the pull_request and push triggers for the Python test workflow.

This prevents the workflow from running on changes unrelated to the source code, tests, or project configuration (e.g., documentation updates).

Additionally, this change adds a `push` trigger to run tests on merges to `main` and `develop`, ensuring the integrity of the primary branches.

GH-62: refactor: Generalize Parquet serialization and statistics logic

Extract the logic for serializing data to Parquet and calculating column statistics into a new, more generic `serialize_dataframe` function in `serialize/parquet.py`.

This new function is now used for serializing both getML DataFrames/Views and prediction results. This refactoring improves code reuse and separation of concerns.

Key changes:
- `serialize_dataframe` accepts callables for saving the Parquet file and retrieving column roles, decoupling it from specific data structures like `getml.DataFrame`.
- `serialize_predictions` now creates a `pyarrow.Table` directly from the numpy array and uses the new `serialize_dataframe` function, avoiding the overhead of creating an intermediate `getml.DataFrame`.
- Moves shared logic out of `serialize_dataframe_or_view.py`, simplifying it significantly.

GH-62: refactor(pydantic): Use `frozen=True` for immutable models

Modernize Pydantic model definitions by replacing the `model_config` dictionary with the `frozen=True` class keyword argument.

This change simplifies the code, makes it more readable, and aligns with current Pydantic v2 best practices for creating immutable models. No functional changes are introduced.

Author: Urfoex

.github/workflows/python-tests.yml

@@ -1,7 +1,20 @@
 name: Python Linting, Formatting, Testing, Coverage
 on:
   pull_request:
-    branches: [main, develop]
+    paths:
+      - 'src/**'
+      - 'tests/**'
+      - '.github/workflows/**'
+      - 'pyproject.toml'
+  push:
+    branches:
+      - main
+      - develop
+    paths:
+      - 'src/**'
+      - 'tests/**'
+      - '.github/workflows/**'
+      - 'pyproject.toml'
 jobs:
   test:
     runs-on: ubuntu-latest

src/getml_io/getml/feature_learning.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 from collections.abc import Set as AbstractSet
-from typing import Annotated, ClassVar, Literal
+from typing import Annotated, Literal
 
 from getml.feature_learning.aggregations.types import (
     FastPropAggregations,
@@ -11,12 +11,10 @@
     CrossEntropyLossType,
     SquareLossType,
 )
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import BaseModel, Field
 
 
-class FastProp(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class FastProp(BaseModel, frozen=True):
     aggregation: AbstractSet[FastPropAggregations]
     delta_t: float
     loss_function: CrossEntropyLossType | SquareLossType | None
@@ -31,9 +29,7 @@ class FastProp(BaseModel):
     type: Literal["fast_prop"] = "fast_prop"
 
 
-class Fastboost(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Fastboost(BaseModel, frozen=True):
     gamma: float
     loss_function: CrossEntropyLossType | SquareLossType | None
     max_depth: int
@@ -48,9 +44,7 @@ class Fastboost(BaseModel):
     type: Literal["fastboost"] = "fastboost"
 
 
-class Multirel(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Multirel(BaseModel, frozen=True):
     aggregation: AbstractSet[MultirelAggregations]
     allow_sets: bool
     delta_t: float
@@ -75,9 +69,7 @@ class Multirel(BaseModel):
     type: Literal["multirel"] = "multirel"
 
 
-class Relboost(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Relboost(BaseModel, frozen=True):
     allow_null_weights: bool
     delta_t: float
     gamma: float
@@ -98,9 +90,7 @@ class Relboost(BaseModel):
     type: Literal["relboost"] = "relboost"
 
 
-class RelMT(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class RelMT(BaseModel, frozen=True):
     allow_avg: bool
     delta_t: float
     gamma: float

src/getml_io/getml/features.py

@@ -1,12 +1,9 @@
 from collections.abc import Mapping
-from typing import ClassVar
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel
 
 
-class Feature(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Feature(BaseModel, frozen=True):
     name: str
     index: int
     target: str

src/getml_io/getml/predictors.py

@@ -1,29 +1,23 @@
 from __future__ import annotations
 
-from typing import Annotated, ClassVar, Literal
+from typing import Annotated, Literal
 
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import BaseModel, Field
 
 
-class LinearRegression(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class LinearRegression(BaseModel, frozen=True):
     learning_rate: float
     reg_lambda: float
     type: Literal["linear_regression"] = "linear_regression"
 
 
-class LogisticRegression(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class LogisticRegression(BaseModel, frozen=True):
     learning_rate: float
     reg_lambda: float
     type: Literal["logistic_regression"] = "logistic_regression"
 
 
-class ScaleGBMClassifier(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class ScaleGBMClassifier(BaseModel, frozen=True):
     colsample_bylevel: float
     colsample_bytree: float
     early_stopping_rounds: int
@@ -41,9 +35,7 @@ class ScaleGBMClassifier(BaseModel):
     type: Literal["scale_gbm_classifier"] = "scale_gbm_classifier"
 
 
-class ScaleGBMRegressor(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class ScaleGBMRegressor(BaseModel, frozen=True):
     colsample_bylevel: float
     colsample_bytree: float
     early_stopping_rounds: int
@@ -61,9 +53,7 @@ class ScaleGBMRegressor(BaseModel):
     type: Literal["scale_gbm_regressor"] = "scale_gbm_regressor"
 
 
-class XGBoostClassifier(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class XGBoostClassifier(BaseModel, frozen=True):
     booster: str
     colsample_bylevel: float
     colsample_bytree: float
@@ -90,9 +80,7 @@ class XGBoostClassifier(BaseModel):
     type: Literal["xgboost_classifier"] = "xgboost_classifier"
 
 
-class XGBoostRegressor(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class XGBoostRegressor(BaseModel, frozen=True):
     booster: str
     colsample_bylevel: float
     colsample_bytree: float

src/getml_io/getml/preprocessors.py

@@ -1,45 +1,35 @@
 from __future__ import annotations
 
 from collections.abc import Set as AbstractSet
-from typing import Annotated, ClassVar, Literal
+from typing import Annotated, Literal
 
 from getml.feature_learning.aggregations.types import MappingAggregations
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import BaseModel, Field
 
 
-class CategoryTrimmer(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class CategoryTrimmer(BaseModel, frozen=True):
     max_num_categories: int
     min_freq: int
     type: Literal["category_trimmer"] = "category_trimmer"
 
 
-class EmailDomain(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class EmailDomain(BaseModel, frozen=True):
     type: Literal["email_domain"] = "email_domain"
 
 
-class Imputation(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Imputation(BaseModel, frozen=True):
     add_dummies: bool
     type: Literal["imputation"] = "imputation"
 
 
-class Mapping(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Mapping(BaseModel, frozen=True):
     aggregation: AbstractSet[MappingAggregations]
     min_freq: int
     multithreading: bool
     type: Literal["mapping"] = "mapping"
 
 
-class Seasonal(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Seasonal(BaseModel, frozen=True):
     disable_year: bool
     disable_month: bool
     disable_weekday: bool
@@ -48,18 +38,14 @@ class Seasonal(BaseModel):
     type: Literal["seasonal"] = "seasonal"
 
 
-class Substring(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Substring(BaseModel, frozen=True):
     begin: int
     length: int
     unit: str
     type: Literal["substring"] = "substring"
 
 
-class TextFieldSplitter(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class TextFieldSplitter(BaseModel, frozen=True):
     type: Literal["text_field_splitter"] = "text_field_splitter"
 
 

src/getml_io/getml/project.py

@@ -1,11 +1,10 @@
 import logging
 from collections.abc import Generator
 from contextlib import contextmanager
-from typing import ClassVar
 
 from getml.data import Container
 from getml.pipeline import Pipeline
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel
 
 from getml_io.getml.exception import (
     PipelineNotFoundError,
@@ -25,12 +24,7 @@
 logger: logging.Logger = logging.getLogger(__name__)
 
 
-class Project(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(
-        arbitrary_types_allowed=True,
-        frozen=True,
-    )
-
+class Project(BaseModel, frozen=True, arbitrary_types_allowed=True):
     name: str
     pipeline: Pipeline
     container: Container

src/getml_io/getml/project_information.py

@@ -1,11 +1,7 @@
-from typing import ClassVar
+from pydantic import BaseModel
 
-from pydantic import BaseModel, ConfigDict
-
-
-class ProjectInformation(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
 
+class ProjectInformation(BaseModel, frozen=True):
     project_name: str
     pipeline_id: str
     container_id: str

src/getml_io/getml/roles.py

@@ -2,10 +2,9 @@
 
 from collections.abc import Sequence
 from enum import Enum
-from typing import ClassVar
 
 from getml.data import roles
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel
 
 
 class Role(str, Enum):
@@ -19,9 +18,7 @@ class Role(str, Enum):
     UNUSED_STRING = roles.unused_string
 
 
-class Roles(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class Roles(BaseModel, frozen=True):
     categorical: Sequence[str]
     join_key: Sequence[str]
     numerical: Sequence[str]

src/getml_io/getml/scores.py

@@ -5,20 +5,20 @@
 from pydantic import BaseModel, Field
 
 
-class _Score(BaseModel):
+class _Score(BaseModel, frozen=True):
     date_time: datetime
     set_used: str
     target: str
 
 
-class ClassificationScore(_Score):
+class ClassificationScore(_Score, frozen=True):
     accuracy: float
     auc: float
     cross_entropy: float
     type: Literal["classification"] = "classification"
 
 
-class RegressionScore(_Score):
+class RegressionScore(_Score, frozen=True):
     mae: float
     rmse: float
     rsquared: float

src/getml_io/metadata/container_information.py

@@ -1,18 +1,14 @@
 from __future__ import annotations
 
-from typing import ClassVar
-
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel
 
 from getml_io.metadata.dataframe_information import (
     DataFrameInformation,
     DataFrameInformationByName,
 )
 
 
-class ContainerInformation(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(frozen=True)
-
+class ContainerInformation(BaseModel, frozen=True):
     id: str
     population: DataFrameInformation | None
     peripheral: DataFrameInformationByName