factrainer-core 0.1.0__tar.gz

factrainer_core-0.1.0/.gitignore

@@ -0,0 +1,173 @@
+notebooks/
+
+# pytest-profiling
+prof/
+
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/

factrainer_core-0.1.0/PKG-INFO

@@ -0,0 +1,7 @@
+Metadata-Version: 2.4
+Name: factrainer-core
+Version: 0.1.0
+Summary: Add your description here
+Author-email: ritsuki1227 <ritsuki1227@gmail.com>
+Requires-Python: >=3.12
+Requires-Dist: factrainer-base

factrainer_core-0.1.0/pyproject.toml

@@ -0,0 +1,18 @@
+[project]
+name = "factrainer-core"
+version = "0.1.0"
+description = "Add your description here"
+readme = "README.md"
+authors = [{ name = "ritsuki1227", email = "ritsuki1227@gmail.com" }]
+requires-python = ">=3.12"
+dependencies = ["factrainer-base"]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/factrainer"]
+
+[tool.uv.sources]
+factrainer-base = { workspace = true }

factrainer_core-0.1.0/src/factrainer/core/__init__.py

@@ -0,0 +1,7 @@
+from .cv.cv import CvMlModel
+from .single import SingleMlModel
+
+__all__ = [
+    "CvMlModel",
+    "SingleMlModel",
+]

factrainer_core-0.1.0/src/factrainer/core/cv/config.py

@@ -0,0 +1,128 @@
+from typing import Any, Self
+
+import numpy as np
+import numpy.typing as npt
+from factrainer.base.config import (
+    BaseLearner,
+    BaseMlModelConfig,
+    BasePredictConfig,
+    BasePredictor,
+    BaseTrainConfig,
+)
+from factrainer.base.dataset import IndexableDataset, Prediction
+from factrainer.base.raw_model import RawModel
+from joblib import Parallel, delayed
+
+from .dataset import IndexedDatasets
+from .raw_model import CvRawModels
+
+
+class CvLearner[T: IndexableDataset, U: RawModel, V: BaseTrainConfig](
+    BaseLearner[IndexedDatasets[T], CvRawModels[U], V]
+):
+    def __init__(
+        self, learner: BaseLearner[T, U, V], n_jobs: int | None = None
+    ) -> None:
+        self._learner = learner
+        self._n_jobs = n_jobs
+
+    def train(
+        self,
+        train_dataset: IndexedDatasets[T],
+        val_dataset: IndexedDatasets[T] | None,
+        config: V,
+    ) -> CvRawModels[U]:
+        if val_dataset is not None:
+            models = Parallel(n_jobs=self.n_jobs)(
+                delayed(self._learner.train)(train.data, val.data, config)
+                for train, val in zip(train_dataset.datasets, val_dataset.datasets)
+            )
+        else:
+            models = Parallel(n_jobs=self.n_jobs)(
+                delayed(self._learner.train)(train.data, None, config)
+                for train in train_dataset.datasets
+            )
+        return CvRawModels(models=models)
+
+    @property
+    def n_jobs(self) -> int | None:
+        return self._n_jobs
+
+    @n_jobs.setter
+    def n_jobs(self, n_jobs: int | None) -> None:
+        self._n_jobs = n_jobs
+
+
+class CvPredictor[T: IndexableDataset, U: RawModel, W: BasePredictConfig](
+    BasePredictor[IndexedDatasets[T], CvRawModels[U], W]
+):
+    def __init__(
+        self, predictor: BasePredictor[T, U, W], n_jobs: int | None = None
+    ) -> None:
+        self._predictor = predictor
+        self._n_jobs = n_jobs
+
+    def predict(
+        self, dataset: IndexedDatasets[T], model: CvRawModels[U], config: W | None
+    ) -> Prediction:
+        y_preds = Parallel(n_jobs=self.n_jobs)(
+            delayed(self._predictor.predict)(_dataset.data, _model, config)
+            for _model, _dataset in zip(model.models, dataset.datasets)
+        )
+        y_oof_pred = self._init_pred(len(dataset), y_preds[0])
+        for y_pred, _dataset in zip(y_preds, dataset.datasets):
+            y_oof_pred[_dataset.index] = y_pred
+        return y_oof_pred
+
+    def _init_pred(self, total_length: int, y_pred: npt.NDArray[Any]) -> Prediction:
+        return np.empty(
+            tuple([total_length] + list(y_pred.shape[1:])), dtype=y_pred.dtype
+        )
+
+    @property
+    def n_jobs(self) -> int | None:
+        return self._n_jobs
+
+    @n_jobs.setter
+    def n_jobs(self, n_jobs: int | None) -> None:
+        self._n_jobs = n_jobs
+
+
+class CvMlModelConfig[
+    T: IndexableDataset,
+    U: RawModel,
+    V: BaseTrainConfig,
+    W: BasePredictConfig,
+](BaseMlModelConfig[IndexedDatasets[T], CvRawModels[U], V, W]):
+    learner: CvLearner[T, U, V]
+    predictor: CvPredictor[T, U, W]
+
+    @classmethod
+    def from_config(
+        cls,
+        config: BaseMlModelConfig[T, U, V, W],
+        n_jobs_train: int | None = None,
+        n_jobs_predict: int | None = None,
+    ) -> Self:
+        return cls(
+            learner=CvLearner(config.learner, n_jobs_train),
+            predictor=CvPredictor(config.predictor, n_jobs_predict),
+            train_config=config.train_config,
+            pred_config=config.pred_config,
+        )
+
+    @property
+    def n_jobs_train(self) -> int | None:
+        return self.learner.n_jobs
+
+    @n_jobs_train.setter
+    def n_jobs_train(self, n_jobs: int | None) -> None:
+        self.learner.n_jobs = n_jobs
+
+    @property
+    def n_jobs_predict(self) -> int | None:
+        return self.predictor.n_jobs
+
+    @n_jobs_predict.setter
+    def n_jobs_predict(self, n_jobs: int | None) -> None:
+        self.predictor.n_jobs = n_jobs

factrainer_core-0.1.0/src/factrainer/core/cv/cv.py

@@ -0,0 +1,78 @@
+from factrainer.base.config import BaseMlModelConfig, BasePredictConfig, BaseTrainConfig
+from factrainer.base.dataset import IndexableDataset, Prediction
+from factrainer.base.raw_model import RawModel
+from sklearn.model_selection._split import _BaseKFold
+
+from ..single import SingleMlModel
+from ..trait import PredictorTrait, TrainerTrait
+from .config import CvMlModelConfig
+from .dataset import IndexedDatasets, SplittedDatasets, SplittedDatasetsIndices
+from .raw_model import CvRawModels
+
+
+class CvMlModel[
+    T: IndexableDataset,
+    U: RawModel,
+    V: BaseTrainConfig,
+    W: BasePredictConfig,
+](TrainerTrait[T, V], PredictorTrait[T, CvRawModels[U], W]):
+    def __init__(
+        self,
+        model_config: BaseMlModelConfig[T, U, V, W],
+        k_fold: _BaseKFold,
+        n_jobs_train: int | None = None,
+        n_jobs_predict: int | None = None,
+    ) -> None:
+        self._cv_model = SingleMlModel(
+            CvMlModelConfig.from_config(model_config, n_jobs_train, n_jobs_predict)
+        )
+        self._k_fold = k_fold
+
+    def train(self, dataset: T) -> None:
+        datasets = SplittedDatasets.create(dataset, self._k_fold)
+        self._cv_indices = datasets.indices
+        self._cv_model.train(datasets.train, datasets.test)
+
+    def predict(self, dataset: T) -> Prediction:
+        datasets = IndexedDatasets.create(dataset, self.cv_indices.test)
+        return self._cv_model.predict(datasets)
+
+    @property
+    def raw_model(self) -> CvRawModels[U]:
+        return self._cv_model.raw_model
+
+    @property
+    def cv_indices(self) -> SplittedDatasetsIndices:
+        return self._cv_indices
+
+    @property
+    def n_jobs_train(self) -> int | None:
+        return self._cv_model.model_config.n_jobs_train  # type: ignore
+
+    @n_jobs_train.setter
+    def n_jobs_train(self, n_jobs: int | None) -> None:
+        self._cv_model.model_config.n_jobs_train = n_jobs  # type: ignore
+
+    @property
+    def n_jobs_predict(self) -> int | None:
+        return self._cv_model.model_config.n_jobs_predict  # type: ignore
+
+    @n_jobs_predict.setter
+    def n_jobs_predict(self, n_jobs: int | None) -> None:
+        self._cv_model.model_config.n_jobs_predict = n_jobs  # type: ignore
+
+    @property
+    def train_config(self) -> V:
+        return self._cv_model.train_config
+
+    @train_config.setter
+    def train_config(self, config: V) -> None:
+        self._cv_model.train_config = config
+
+    @property
+    def pred_config(self) -> W | None:
+        return self._cv_model.pred_config
+
+    @pred_config.setter
+    def pred_config(self, config: W | None) -> None:
+        self._cv_model.pred_config = config

factrainer_core-0.1.0/src/factrainer/core/cv/dataset.py

@@ -0,0 +1,99 @@
+from collections.abc import Sequence
+from typing import Self
+
+from factrainer.base.dataset import (
+    BaseDataset,
+    IndexableDataset,
+    RowIndex,
+    RowIndices,
+)
+from sklearn.model_selection._split import _BaseKFold
+
+
+class IndexedDataset[T: IndexableDataset](BaseDataset):
+    index: RowIndex
+    data: T
+
+    def __len__(self) -> int:
+        return len(self.index)
+
+
+class IndexedDatasets[T: IndexableDataset](BaseDataset):
+    datasets: Sequence[IndexedDataset[T]]
+
+    def __len__(self) -> int:
+        return sum([len(dataset) for dataset in self.datasets])
+
+    @classmethod
+    def create(cls, dataset: T, k_fold: _BaseKFold | RowIndices) -> Self:
+        if isinstance(k_fold, _BaseKFold):
+            raise NotImplementedError
+        return cls(
+            datasets=[
+                IndexedDataset(index=index, data=dataset[index]) for index in k_fold
+            ]
+        )
+
+    @property
+    def indices(self) -> RowIndices:
+        return [dataset.index for dataset in self.datasets]
+
+
+class SplittedDataset[T: IndexableDataset](BaseDataset):
+    train: IndexedDataset[T]
+    val: IndexedDataset[T] | None
+    test: IndexedDataset[T]
+
+
+class SplittedDatasetsIndices(BaseDataset):
+    train: RowIndices
+    val: RowIndices | None
+    test: RowIndices
+
+
+class SplittedDatasets[T: IndexableDataset](BaseDataset):
+    datasets: Sequence[SplittedDataset[T]]
+
+    @property
+    def train(self) -> IndexedDatasets[T]:
+        return IndexedDatasets(datasets=[dataset.train for dataset in self.datasets])
+
+    @property
+    def val(self) -> IndexedDatasets[T] | None:
+        vals = []
+        for dataset in self.datasets:
+            if dataset.val is None:
+                return None
+            vals.append(dataset.val)
+        return IndexedDatasets(datasets=vals)
+
+    @property
+    def test(self) -> IndexedDatasets[T]:
+        return IndexedDatasets(datasets=[dataset.test for dataset in self.datasets])
+
+    @property
+    def indices(self) -> SplittedDatasetsIndices:
+        return SplittedDatasetsIndices(
+            train=self.train.indices,
+            val=self.val.indices if self.val is not None else None,
+            test=self.test.indices,
+        )
+
+    @classmethod
+    def create(
+        cls, dataset: T, k_fold: _BaseKFold, share_holdouts: bool = True
+    ) -> Self:
+        datasets = []
+        for train_index, val_index in dataset.get_index(k_fold):
+            if share_holdouts:
+                test_index = val_index
+            else:
+                raise NotImplementedError
+            datasets.append(
+                SplittedDataset(
+                    train=IndexedDataset(index=train_index, data=dataset[train_index]),
+                    val=IndexedDataset(index=val_index, data=dataset[val_index]),
+                    test=IndexedDataset(index=test_index, data=dataset[test_index]),
+                )
+            )
+        return cls(datasets=datasets)

factrainer_core-0.1.0/src/factrainer/core/cv/raw_model.py

@@ -0,0 +1,7 @@
+from collections.abc import Sequence
+
+from factrainer.base.raw_model import RawModel
+
+
+class CvRawModels[U: RawModel](RawModel):
+    models: Sequence[U]

factrainer_core-0.1.0/src/factrainer/core/single.py

@@ -0,0 +1,55 @@
+from factrainer.base.config import (
+    BaseMlModelConfig,
+    BasePredictConfig,
+    BaseTrainConfig,
+)
+from factrainer.base.dataset import BaseDataset, Prediction
+from factrainer.base.raw_model import RawModel
+
+from .trait import (
+    PredictorTrait,
+    ValidatableTrainerTrait,
+)
+
+
+class SingleMlModel[
+    T: BaseDataset,
+    U: RawModel,
+    V: BaseTrainConfig,
+    W: BasePredictConfig,
+](ValidatableTrainerTrait[T, V], PredictorTrait[T, U, W]):
+    def __init__(
+        self,
+        model_config: BaseMlModelConfig[T, U, V, W],
+    ) -> None:
+        self.model_config = model_config
+
+    def train(self, train_dataset: T, val_dataset: T | None = None) -> None:
+        self._model = self.model_config.learner.train(
+            train_dataset, val_dataset, self.model_config.train_config
+        )
+
+    def predict(self, dataset: T) -> Prediction:
+        return self.model_config.predictor.predict(
+            dataset, self.raw_model, self.model_config.pred_config
+        )
+
+    @property
+    def raw_model(self) -> U:
+        return self._model
+
+    @property
+    def train_config(self) -> V:
+        return self.model_config.train_config
+
+    @train_config.setter
+    def train_config(self, config: V) -> None:
+        self.model_config.train_config = config
+
+    @property
+    def pred_config(self) -> W | None:
+        return self.model_config.pred_config
+
+    @pred_config.setter
+    def pred_config(self, config: W | None) -> None:
+        self.model_config.pred_config = config

factrainer_core-0.1.0/src/factrainer/core/trait.py

@@ -0,0 +1,58 @@
+from abc import ABC, abstractmethod
+
+from factrainer.base.config import BasePredictConfig, BaseTrainConfig
+from factrainer.base.dataset import BaseDataset, Prediction
+from factrainer.base.raw_model import RawModel
+
+
+class PredictorTrait[T: BaseDataset, U: RawModel, W: BasePredictConfig](ABC):
+    @abstractmethod
+    def predict(self, dataset: T) -> Prediction:
+        raise NotImplementedError
+
+    @property
+    @abstractmethod
+    def raw_model(self) -> U:
+        raise NotImplementedError
+
+    @property
+    @abstractmethod
+    def pred_config(self) -> W | None:
+        raise NotImplementedError
+
+    @pred_config.setter
+    @abstractmethod
+    def pred_config(self, config: W | None) -> None:
+        raise NotImplementedError
+
+
+class TrainerTrait[T: BaseDataset, V: BaseTrainConfig](ABC):
+    @abstractmethod
+    def train(self, dataset: T) -> None:
+        raise NotImplementedError
+
+    @property
+    @abstractmethod
+    def train_config(self) -> V:
+        raise NotImplementedError
+
+    @train_config.setter
+    @abstractmethod
+    def train_config(self, config: V) -> None:
+        raise NotImplementedError
+
+
+class ValidatableTrainerTrait[T: BaseDataset, V: BaseTrainConfig](ABC):
+    @abstractmethod
+    def train(self, train_dataset: T, val_dataset: T | None = None) -> None:
+        raise NotImplementedError
+
+    @property
+    @abstractmethod
+    def train_config(self) -> V:
+        raise NotImplementedError
+
+    @train_config.setter
+    @abstractmethod
+    def train_config(self, config: V) -> None:
+        raise NotImplementedError