sdg-hub 0.1.3__py3-none-any.whl → 0.2.0__py3-none-any.whl

sdg_hub/core/blocks/deprecated_blocks/rename_columns.py

@@ -0,0 +1,88 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Deprecated RenameColumns for backwards compatibility.
+
+This module provides a deprecated wrapper around RenameColumnsBlock
+to maintain backwards compatibility with existing code and configurations.
+"""
+
+# Standard
+from typing import Any
+import warnings
+
+# Third Party
+from datasets import Dataset
+
+# Local
+from ...utils.logger_config import setup_logger
+from ..base import BaseBlock
+from ..registry import BlockRegistry
+from ..transform import RenameColumnsBlock
+
+logger = setup_logger(__name__)
+
+
+@BlockRegistry.register(
+    "RenameColumns",
+    "deprecated",
+    "DEPRECATED: Use RenameColumnsBlock instead. Renames columns in a dataset according to a mapping dictionary",
+)
+class RenameColumns(BaseBlock):
+    """DEPRECATED: Block for renaming columns in a dataset.
+
+    This block is deprecated and maintained only for backwards compatibility.
+    Please use RenameColumnsBlock instead.
+
+    This block renames columns in a dataset according to a mapping dictionary,
+    where keys are existing column names and values are new column names.
+    """
+
+    def __init__(
+        self,
+        block_name: str,
+        columns_map: dict[str, str],
+    ) -> None:
+        """Initialize the deprecated RenameColumns.
+
+        Parameters
+        ----------
+        block_name : str
+            Name of the block.
+        columns_map : Dict[str, str]
+            Dictionary mapping existing column names to new column names.
+            Keys are existing column names, values are new column names.
+        """
+        # Issue deprecation warning
+        warnings.warn(
+            "RenameColumns is deprecated and will be removed in a future version. "
+            "Please use RenameColumnsBlock instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        # Map old signature to new signature
+        super().__init__(
+            block_name=block_name,
+            input_cols=columns_map,
+            output_cols=[],
+        )
+
+        # Create the new block instance with mapped parameters
+        self._new_block = RenameColumnsBlock(
+            block_name=block_name,
+            input_cols=columns_map,
+        )
+
+    def generate(self, samples: Dataset, **kwargs: Any) -> Dataset:
+        """Generate dataset with renamed columns using the new RenameColumnsBlock.
+
+        Parameters
+        ----------
+        samples : Dataset
+            The input dataset to rename columns in.
+
+        Returns
+        -------
+        Dataset
+            The dataset with renamed columns.
+        """
+        return self._new_block.generate(samples, **kwargs)

sdg_hub/core/blocks/deprecated_blocks/sample_populator.py

@@ -0,0 +1,58 @@
+# SPDX-License-Identifier: Apache-2.0
+"""DEPRECATED: SamplePopulatorBlock for backward compatibility.
+
+This module provides a deprecated stub for SamplePopulatorBlock.
+This block is deprecated and will be replaced with a router block.
+"""
+
+# Standard
+from typing import Any
+import warnings
+
+# Third Party
+from datasets import Dataset
+
+# Local
+from ...utils.logger_config import setup_logger
+from ..base import BaseBlock
+from ..registry import BlockRegistry
+
+logger = setup_logger(__name__)
+
+
+@BlockRegistry.register(
+    "SamplePopulatorBlock",
+    "deprecated",
+    "DEPRECATED: Use a router block instead. Populates dataset with data from configuration files",
+)
+class SamplePopulatorBlock(BaseBlock):
+    """DEPRECATED: Block for populating dataset with data from configuration files.
+
+    .. deprecated::
+        This block is deprecated and will be replaced with a router block.
+    """
+
+    def __init__(
+        self,
+        block_name: str,
+        config_paths: list[str],
+        column_name: str,
+        post_fix: str = "",
+        **batch_kwargs: dict[str, Any],
+    ) -> None:
+        warnings.warn(
+            "SamplePopulatorBlock is deprecated and will be replaced with a router block.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        # Initialize with dummy values for BaseBlock validation
+        super().__init__(
+            block_name=block_name, input_cols=[column_name], output_cols=[column_name]
+        )
+
+    def generate(self, samples: Dataset, **kwargs: Any) -> Dataset:
+        """Generate method - raises error as block is deprecated."""
+        raise NotImplementedError(
+            "SamplePopulatorBlock is deprecated and will be replaced with a router block."
+        )

sdg_hub/core/blocks/deprecated_blocks/selector.py

@@ -0,0 +1,97 @@
+# SPDX-License-Identifier: Apache-2.0
+"""DEPRECATED: SelectorBlock for backward compatibility.
+
+This module provides a deprecated wrapper for the old SelectorBlock interface.
+Use transform.IndexBasedMapperBlock instead.
+"""
+
+# Standard
+from typing import Any
+import warnings
+
+# Third Party
+from datasets import Dataset
+
+# Local
+from ...utils.logger_config import setup_logger
+from ..base import BaseBlock
+from ..registry import BlockRegistry
+from ..transform.index_based_mapper import IndexBasedMapperBlock
+
+logger = setup_logger(__name__)
+
+
+@BlockRegistry.register(
+    "SelectorBlock",
+    "deprecated",
+    "DEPRECATED: Use IndexBasedMapperBlock instead. Selects and maps values from one column to another",
+)
+class SelectorBlock(BaseBlock):
+    """DEPRECATED: Block for selecting and mapping values from one column to another.
+
+    .. deprecated::
+        Use `sdg_hub.blocks.transform.IndexBasedMapperBlock` instead.
+        This class will be removed in a future version.
+
+    This block uses a mapping dictionary to select values from one column and
+    store them in a new output column based on a choice column's value.
+
+    Parameters
+    ----------
+    block_name : str
+        Name of the block.
+    choice_map : Dict[str, str]
+        Dictionary mapping choice values to column names.
+    choice_col : str
+        Name of the column containing choice values.
+    output_col : str
+        Name of the column to store selected values.
+    **batch_kwargs : Dict[str, Any]
+        Additional keyword arguments for batch processing.
+    """
+
+    def __init__(
+        self,
+        block_name: str,
+        choice_map: dict[str, str],
+        choice_col: str,
+        output_col: str,
+        **batch_kwargs: dict[str, Any],
+    ) -> None:
+        warnings.warn(
+            "SelectorBlock is deprecated. Use sdg_hub.blocks.transform.IndexBasedMapperBlock instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        # Initialize with dummy values for BaseBlock validation
+        # We need all columns referenced in choice_map as input, plus the choice column
+        all_input_cols = list(choice_map.values()) + [choice_col]
+
+        super().__init__(
+            block_name=block_name, input_cols=all_input_cols, output_cols=[output_col]
+        )
+
+        # Create the new implementation
+        self._impl = IndexBasedMapperBlock(
+            block_name=block_name,
+            input_cols=all_input_cols,
+            output_cols=[output_col],
+            choice_map=choice_map,
+            choice_cols=[choice_col],
+        )
+
+    def generate(self, samples: Dataset) -> Dataset:
+        """Generate a new dataset with selected values.
+
+        Parameters
+        ----------
+        samples : Dataset
+            Input dataset to process.
+
+        Returns
+        -------
+        Dataset
+            Dataset with selected values stored in output column.
+        """
+        return self._impl.generate(samples)

sdg_hub/core/blocks/deprecated_blocks/set_to_majority_value.py

@@ -0,0 +1,88 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Deprecated SetToMajorityValue for backwards compatibility.
+
+This module provides a deprecated wrapper around UniformColumnValueSetter
+to maintain backwards compatibility with existing code and configurations.
+"""
+
+# Standard
+from typing import Any
+import warnings
+
+# Third Party
+from datasets import Dataset
+
+# Local
+from ...utils.logger_config import setup_logger
+from ..base import BaseBlock
+from ..registry import BlockRegistry
+from ..transform import UniformColumnValueSetter
+
+logger = setup_logger(__name__)
+
+
+@BlockRegistry.register(
+    "SetToMajorityValue",
+    "deprecated",
+    "DEPRECATED: Use UniformColumnValueSetter with reduction_strategy='mode' instead. Sets all values in a column to the most frequent value",
+)
+class SetToMajorityValue(BaseBlock):
+    """DEPRECATED: Block for setting all values in a column to the most frequent value.
+
+    This block is deprecated and maintained only for backwards compatibility.
+    Please use UniformColumnValueSetter with reduction_strategy='mode' instead.
+
+    This block finds the most common value (mode) in a specified column and
+    replaces all values in that column with this majority value.
+    """
+
+    def __init__(
+        self,
+        block_name: str,
+        col_name: str,
+    ) -> None:
+        """Initialize the deprecated SetToMajorityValue.
+
+        Parameters
+        ----------
+        block_name : str
+            Name of the block.
+        col_name : str
+            Name of the column to set to majority value.
+        """
+        # Issue deprecation warning
+        warnings.warn(
+            "SetToMajorityValue is deprecated and will be removed in a future version. "
+            "Please use UniformColumnValueSetter with reduction_strategy='mode' instead.",
+            DeprecationWarning,
+            stacklevel=2,
+        )
+
+        # Map old signature to new signature
+        super().__init__(
+            block_name=block_name,
+            input_cols=[col_name],
+            output_cols=[],
+        )
+
+        # Create the new block instance with mapped parameters
+        self._new_block = UniformColumnValueSetter(
+            block_name=block_name,
+            input_cols=[col_name],
+            reduction_strategy="mode",
+        )
+
+    def generate(self, samples: Dataset, **kwargs: Any) -> Dataset:
+        """Generate dataset with column set to majority value using UniformColumnValueSetter.
+
+        Parameters
+        ----------
+        samples : Dataset
+            The input dataset to process.
+
+        Returns
+        -------
+        Dataset
+            The dataset with specified column set to its majority value.
+        """
+        return self._new_block.generate(samples, **kwargs)

sdg_hub/core/blocks/evaluation/__init__.py

@@ -0,0 +1,9 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Evaluation blocks for SDG Hub."""
+
+# Local
+from .evaluate_faithfulness_block import EvaluateFaithfulnessBlock
+from .evaluate_relevancy_block import EvaluateRelevancyBlock
+from .verify_question_block import VerifyQuestionBlock
+
+__all__ = ["EvaluateFaithfulnessBlock", "EvaluateRelevancyBlock", "VerifyQuestionBlock"]