Named Entity Recognition

The NER task identifies and classifies named entities in text.

Usage

Simple List of Entities

You can provide a simple list of entity types to extract.

from sieves import tasks

task = tasks.NER(
    entities=["PERSON", "ORGANIZATION", "LOCATION"],
    model=model,
)

Entities with Descriptions (Recommended)

Providing descriptions for each entity type helps the model understand exactly what you are looking for.

task = tasks.NER(
    entities={
        "PERSON": "Names of people.",
        "ORGANIZATION": "Companies, agencies, institutions.",
    },
    model=model,
)

Results

The NER task returns a unified Result object (an alias for Entities) containing a list of Entity objects and the source text.

Each entity includes a confidence score: - GLiNER2: Always present and derived from logits. - LLMs: Self-reported and may be None if not provided by the model.

class Result(Entities):
    """Result of a named-entity recognition (NER) task. Contains the extracted entities and the source text."""

    pass

Evaluation

Performance of the NER task can be measured using the .evaluate() method.

• Metric: Corpus-wide Micro-F1 Score (F1). Entities are matched based on their text span (start/end offsets) and type. True Positives, False Positives, and False Negatives are accumulated across the entire dataset.
• Requirement: Each document must have ground-truth entities stored in doc.gold[task_id].

report = task.evaluate(docs)
print(f"NER F1-Score: {report.metrics['F1']}")

Ground Truth Formats

Ground truth has to be specified in doc.meta using Result instances.

---

Named‑Entity Recognition (NER) predictive task.

NER

Bases: PredictiveTask[TaskPromptSignature, TaskResult, _TaskBridge]

Extract named entities from text using various model wrappers.

Source code in sieves/tasks/predictive/ner/core.py

class NER(PredictiveTask[TaskPromptSignature, TaskResult, _TaskBridge]):
    """Extract named entities from text using various model wrappers."""

    def __init__(
        self,
        entities: Sequence[str] | dict[str, str] | None = None,
        model: TaskModel | None = None,
        task_id: str | None = None,
        include_meta: bool = True,
        batch_size: int = -1,
        prompt_instructions: str | None = None,
        fewshot_examples: Sequence[FewshotExample] = (),
        model_settings: ModelSettings = ModelSettings(),
        condition: Callable[[Doc], bool] | None = None,
    ) -> None:
        """Initialize NER task.

        :param entities: Entity types to extract. Supports two formats:
            - List format: `["PERSON", "LOCATION", "ORGANIZATION"]`
            - Dict format: `{"PERSON": "Names of people", "LOCATION": "Geographic locations"}`
            The dict format allows you to provide descriptions that help the model better identify entities.
            When using GliNER, descriptions are passed directly to the model for improved recognition.
            If None, defaults to `["PERSON", "LOCATION", "ORGANIZATION"]`.
        :param model: Model to use.
        :param task_id: Task ID.
        :param include_meta: Whether to include meta information generated by the task.
        :param batch_size: Batch size to use for inference. Use -1 to process all documents at once.
        :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
        :param fewshot_examples: Few-shot examples.
        :param model_settings: Settings for structured generation.
        :param condition: Optional callable that determines whether to process each document.
        """
        if entities is None:
            entities = ["PERSON", "LOCATION", "ORGANIZATION"]

        if isinstance(entities, dict):
            self._entities = list(entities.keys())
            self._entity_descriptions = entities
        else:
            self._entities = list(entities)
            self._entity_descriptions = {}

        self._entities_param = entities

        super().__init__(
            model=model,
            task_id=task_id,
            include_meta=include_meta,
            batch_size=batch_size,
            overwrite=False,
            prompt_instructions=prompt_instructions,
            fewshot_examples=fewshot_examples,
            model_settings=model_settings,
            condition=condition,
        )

    @property
    @override
    def fewshot_example_type(self) -> type[FewshotExample]:
        return FewshotExample

    @property
    @override
    def prompt_signature(self) -> type[pydantic.BaseModel]:
        """Return the unified Pydantic prompt signature for this task.

        :return: Unified Pydantic prompt signature.
        """
        # Create a dynamic entity model with Literal for the entity types.
        EntityTypes = Literal[*(tuple(self._entities))] if self._entities else str  # type: ignore[valid-type]

        DynamicEntity = pydantic.create_model(
            "NEREntity",
            __doc__="Extracted named entity with its context and type.",
            text=(str, pydantic.Field(description="The specific text segment identified as an entity.")),
            context=(str, pydantic.Field(description="The surrounding text providing context for the entity.")),
            entity_type=(
                EntityTypes,
                pydantic.Field(description="The category or type of the entity (e.g., PERSON, ORGANIZATION)."),
            ),
            score=(
                float | None,
                pydantic.Field(
                    default=None,
                    description="Provide a confidence score for the entity identification, between 0 and 1.",
                ),
            ),
            __base__=pydantic.BaseModel,
        )

        return pydantic.create_model(
            "NEROutput",
            __doc__="Result of named-entity recognition. Contains a list of extracted entities.",
            entities=(
                list[DynamicEntity],  # type: ignore[valid-type]
                pydantic.Field(..., description="List of extracted named entities."),
            ),
        )

    @property
    @override
    def metric(self) -> str:
        return "F1"

    @override
    def _compute_metrics(self, truths: list[Any], preds: list[Any], judge: dspy.LM | None = None) -> dict[str, float]:
        """Compute corpus-level metrics.

        :param truths: List of ground truths.
        :param preds: List of predictions.
        :param judge: Optional DSPy LM instance to use as judge for generative tasks.
        :return: Dictionary of metrics.
        """
        tp = 0
        fp = 0
        fn = 0

        for gold, pred in zip(truths, preds):
            if gold is None:
                true_entities = set()
            else:
                # Convert to DSPy format first to reuse logic or do it directly.
                assert isinstance(gold, TaskResult)
                true_entities = {(e.text, e.entity_type) for e in gold.entities}

            if pred is None:
                pred_entities = set()
            else:
                if hasattr(pred, "entities"):
                    pred_entities = {(e.text, e.entity_type) for e in pred.entities}
                elif isinstance(pred, dict) and "entities" in pred:
                    pred_entities = {(e["text"], e["entity_type"]) for e in pred.get("entities", [])}
                else:
                    pred_entities = set()

            tp += len(true_entities & pred_entities)
            fp += len(pred_entities - true_entities)
            fn += len(true_entities - pred_entities)

        precision = tp / (tp + fp) if (tp + fp) > 0 else 0.0
        recall = tp / (tp + fn) if (tp + fn) > 0 else 0.0
        f1 = 2 * precision * recall / (precision + recall) if (precision + recall) > 0 else 0.0

        return {self.metric: f1}

    @override
    def _init_bridge(self, model_type: ModelType) -> _TaskBridge:
        if model_type == ModelType.gliner:
            from sieves.model_wrappers import gliner_

            return GliNERBridge(
                task_id=self._task_id,
                prompt_instructions=self._custom_prompt_instructions,
                prompt_signature=self.prompt_signature,
                model_settings=self._model_settings,
                inference_mode=gliner_.InferenceMode.entities,
            )

        bridge_types = {
            ModelType.langchain: PydanticNER,
            ModelType.outlines: PydanticNER,
            ModelType.dspy: DSPyNER,
        }

        try:
            bridge_class = bridge_types[model_type]
            result = bridge_class(
                task_id=self._task_id,
                prompt_instructions=self._custom_prompt_instructions,
                entities=self._entities_param,
                model_settings=self._model_settings,
                prompt_signature=self.prompt_signature,
                model_type=model_type,
                fewshot_examples=self._fewshot_examples,
            )
            return result  # type: ignore[return-value]
        except KeyError as err:
            raise KeyError(f"Model type {model_type} is not supported by {self.__class__.__name__}.") from err

    @staticmethod
    @override
    def supports() -> set[ModelType]:
        return {
            ModelType.langchain,
            ModelType.dspy,
            ModelType.outlines,
            ModelType.gliner,
        }

    @override
    def _validate_fewshot_examples(self) -> None:
        for fs_example in self._fewshot_examples or []:
            assert isinstance(fs_example, FewshotExample)
            for entity in fs_example.entities:
                if entity.entity_type not in self._entities:
                    raise ValueError(f"Entity {entity.entity_type} not in {self._entities}.")

    @override
    @property
    def _state(self) -> dict[str, Any]:
        # Store entities as dict if descriptions exist, else as list
        entities_state = self._entity_descriptions if self._entity_descriptions else self._entities
        return {
            **super()._state,
            "entities": entities_state,
        }

    @override
    def to_hf_dataset(self, docs: Iterable[Doc], threshold: float | None = None) -> datasets.Dataset:
        # Define metadata and features for the dataset
        features = datasets.Features(
            {
                "text": datasets.Value("string"),
                "entities": datasets.Sequence(
                    datasets.Features(
                        {
                            "text": datasets.Value("string"),
                            "start": datasets.Value("int32"),
                            "end": datasets.Value("int32"),
                            "entity_type": datasets.Value("string"),
                            "score": datasets.Value("float32"),
                        }
                    )
                ),
            }
        )

        info = datasets.DatasetInfo(
            description=f"Named Entity Recognition dataset with entity types {self._entities}. Generated with sieves "
            f"v{Config.get_version()}.",
            features=features,
        )

        # Fetch data used for generating dataset
        try:
            data: list[tuple[str, list[dict[str, Any]]]] = []
            for doc in docs:
                if self._task_id not in doc.results:
                    raise KeyError(f"Document does not have results for task ID {self._task_id}")

                # Get the entities from the document results
                result = doc.results[self._task_id].entities
                entities: list[dict[str, Any]] = []

                # List format (could be list of dictionaries or other entities)
                for entity in result:
                    assert hasattr(entity, "text")
                    assert hasattr(entity, "start")
                    assert hasattr(entity, "end")
                    assert hasattr(entity, "entity_type")

                    entities.append(
                        {
                            "text": entity.text,
                            "start": entity.start,
                            "end": entity.end,
                            "entity_type": entity.entity_type,
                            "score": getattr(entity, "score", None),
                        }
                    )

                data.append((doc.text or "", entities))

        except KeyError as err:
            raise KeyError(f"Not all documents have results for this task with ID {self._task_id}") from err

        def generate_data() -> Iterable[dict[str, Any]]:
            """Yield results as dicts.

            :return: Results as dicts.
            """
            for text, entities in data:
                yield {"text": text, "entities": entities}

        # Create dataset
        return datasets.Dataset.from_generator(generate_data, features=features, info=info)

    @override
    def distill(
        self,
        base_model_id: str,
        framework: DistillationFramework,
        data: datasets.Dataset | Sequence[Doc],
        output_path: Path | str,
        val_frac: float,
        init_kwargs: dict[str, Any] | None = None,
        train_kwargs: dict[str, Any] | None = None,
        seed: int | None = None,
    ) -> None:
        raise NotImplementedError

    @override
    def _evaluate_dspy_example(self, truth: dspy.Example, pred: dspy.Prediction, trace: Any, model: dspy.LM) -> float:
        # Compute entity-level F1 score based on (text, entity_type) pairs
        true_entities = {(e["text"], e["entity_type"]) for e in truth["entities"]}
        pred_entities = {(e["text"], e["entity_type"]) for e in pred.get("entities", [])}

        if not true_entities:
            return 1.0 if not pred_entities else 0.0

        precision = len(true_entities & pred_entities) / len(pred_entities) if pred_entities else 0
        recall = len(true_entities & pred_entities) / len(true_entities)
        return 2 * precision * recall / (precision + recall) if (precision + recall) > 0 else 0

fewshot_examples property

Return few-shot examples.

Returns:

Type	Description
Sequence[FewshotExample]	Few-shot examples.

id property

Return task ID.

Used by pipeline for results and dependency management.

Returns:

Type	Description
str	Task ID.

prompt_signature property

Return the unified Pydantic prompt signature for this task.

Returns:

Type	Description
type[BaseModel]	Unified Pydantic prompt signature.

prompt_signature_description property

Return prompt signature description.

Returns:

Type	Description
str | None	Prompt signature description.

prompt_template property

Return prompt template.

Returns:

Type	Description
str	Prompt template.

__add__(other)

Chain this task with another task or pipeline using the + operator.

This returns a new Pipeline that executes this task first, followed by the task(s) in other. The original task(s)/pipeline are not mutated.

Cache semantics: - If other is a Pipeline, the resulting pipeline adopts other's use_cache setting (because the left-hand side is a single task). - If other is a Task, the resulting pipeline defaults to use_cache=True.

Parameters:

Name	Type	Description	Default
other	Task | Pipeline	A Task or Pipeline to execute after this task.	required

Returns:

Type	Description
Pipeline	A new Pipeline representing the chained execution.

Raises:

Type	Description
TypeError	If other is not a Task or Pipeline.

Source code in sieves/tasks/core.py

def __add__(self, other: Task | Pipeline) -> Pipeline:
    """Chain this task with another task or pipeline using the ``+`` operator.

    This returns a new ``Pipeline`` that executes this task first, followed by the
    task(s) in ``other``. The original task(s)/pipeline are not mutated.

    Cache semantics:
    - If ``other`` is a ``Pipeline``, the resulting pipeline adopts ``other``'s
      ``use_cache`` setting (because the left-hand side is a single task).
    - If ``other`` is a ``Task``, the resulting pipeline defaults to ``use_cache=True``.

    :param other: A ``Task`` or ``Pipeline`` to execute after this task.
    :return: A new ``Pipeline`` representing the chained execution.
    :raises TypeError: If ``other`` is not a ``Task`` or ``Pipeline``.
    """
    # Lazy import to avoid circular dependency at module import time.
    from sieves.pipeline import Pipeline

    if isinstance(other, Pipeline):
        return Pipeline(tasks=[self, *other.tasks], use_cache=other.use_cache)

    if isinstance(other, Task):
        return Pipeline(tasks=[self, other])

    raise TypeError(f"Cannot chain Task with {type(other).__name__}")

__call__(docs)

Execute task with conditional logic.

Checks the condition for each document without materializing all docs upfront. Passes all documents that pass the condition to _call() for proper batching. Documents that fail the condition have results[task_id] set to None.

Parameters:

Name	Type	Description	Default
docs	Iterable[Doc]	Docs to process.	required

Returns:

Type	Description
Iterable[Doc]	Processed docs (in original order).

Source code in sieves/tasks/core.py

def __call__(self, docs: Iterable[Doc]) -> Iterable[Doc]:
    """Execute task with conditional logic.

    Checks the condition for each document without materializing all docs upfront.
    Passes all documents that pass the condition to _call() for proper batching.
    Documents that fail the condition have results[task_id] set to None.

    :param docs: Docs to process.
    :return: Processed docs (in original order).
    """
    docs = iter(docs) if not isinstance(docs, Iterator) else docs

    # Materialize docs in batches. This doesn't incur additional memory overhead, as docs are materialized in
    # batches downstream anyway.
    batch_size = self._batch_size if self._batch_size > 0 else sys.maxsize
    while docs_batch := [doc for doc in itertools.islice(docs, batch_size)]:
        # First pass: determine which docs pass the condition by index.
        passing_indices: set[int] = {
            idx for idx, doc in enumerate(docs_batch) if self._condition is None or self._condition(doc)
        }

        # Process all passing docs in one batch.
        processed = self._call(d for i, d in enumerate(docs_batch) if i in passing_indices)
        processed_iter = iter(processed) if not isinstance(processed, Iterator) else processed

        # Iterate through original docs in order and yield results.
        for idx, doc in enumerate(docs_batch):
            if idx in passing_indices:
                # Doc passed condition - use processed result.
                yield next(processed_iter)
            else:
                # Doc failed condition - set `None` result and yield original.
                doc.results[self.id] = None
                yield doc

__init__(entities=None, model=None, task_id=None, include_meta=True, batch_size=-1, prompt_instructions=None, fewshot_examples=(), model_settings=ModelSettings(), condition=None)

Initialize NER task.

Parameters:

Name	Type	Description	Default
entities	Sequence[str] | dict[str, str] | None	Entity types to extract. Supports two formats: - List format: ["PERSON", "LOCATION", "ORGANIZATION"] - Dict format: {"PERSON": "Names of people", "LOCATION": "Geographic locations"} The dict format allows you to provide descriptions that help the model better identify entities. When using GliNER, descriptions are passed directly to the model for improved recognition. If None, defaults to ["PERSON", "LOCATION", "ORGANIZATION"].	None
model	TaskModel | None	Model to use.	None
task_id	str | None	Task ID.	None
include_meta	bool	Whether to include meta information generated by the task.	True
batch_size	int	Batch size to use for inference. Use -1 to process all documents at once.	-1
prompt_instructions	str | None	Custom prompt instructions. If None, default instructions are used.	None
fewshot_examples	Sequence[FewshotExample]	Few-shot examples.	()
model_settings	ModelSettings	Settings for structured generation.	ModelSettings()
condition	Callable[[Doc], bool] | None	Optional callable that determines whether to process each document.	None

Source code in sieves/tasks/predictive/ner/core.py

def __init__(
    self,
    entities: Sequence[str] | dict[str, str] | None = None,
    model: TaskModel | None = None,
    task_id: str | None = None,
    include_meta: bool = True,
    batch_size: int = -1,
    prompt_instructions: str | None = None,
    fewshot_examples: Sequence[FewshotExample] = (),
    model_settings: ModelSettings = ModelSettings(),
    condition: Callable[[Doc], bool] | None = None,
) -> None:
    """Initialize NER task.

    :param entities: Entity types to extract. Supports two formats:
        - List format: `["PERSON", "LOCATION", "ORGANIZATION"]`
        - Dict format: `{"PERSON": "Names of people", "LOCATION": "Geographic locations"}`
        The dict format allows you to provide descriptions that help the model better identify entities.
        When using GliNER, descriptions are passed directly to the model for improved recognition.
        If None, defaults to `["PERSON", "LOCATION", "ORGANIZATION"]`.
    :param model: Model to use.
    :param task_id: Task ID.
    :param include_meta: Whether to include meta information generated by the task.
    :param batch_size: Batch size to use for inference. Use -1 to process all documents at once.
    :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
    :param fewshot_examples: Few-shot examples.
    :param model_settings: Settings for structured generation.
    :param condition: Optional callable that determines whether to process each document.
    """
    if entities is None:
        entities = ["PERSON", "LOCATION", "ORGANIZATION"]

    if isinstance(entities, dict):
        self._entities = list(entities.keys())
        self._entity_descriptions = entities
    else:
        self._entities = list(entities)
        self._entity_descriptions = {}

    self._entities_param = entities

    super().__init__(
        model=model,
        task_id=task_id,
        include_meta=include_meta,
        batch_size=batch_size,
        overwrite=False,
        prompt_instructions=prompt_instructions,
        fewshot_examples=fewshot_examples,
        model_settings=model_settings,
        condition=condition,
    )

deserialize(config, **kwargs) classmethod

Generate PredictiveTask instance from config.

Parameters:

Name	Type	Description	Default
config	Config	Config to generate instance from.	required
kwargs	dict[str, Any]	Values to inject into loaded config.	{}

Returns:

Type	Description
PredictiveTask[TaskPromptSignature, TaskResult, TaskBridge]	Deserialized PredictiveTask instance.

Source code in sieves/tasks/predictive/core.py

@classmethod
def deserialize(
    cls, config: Config, **kwargs: dict[str, Any]
) -> PredictiveTask[TaskPromptSignature, TaskResult, TaskBridge]:
    """Generate PredictiveTask instance from config.

    :param config: Config to generate instance from.
    :param kwargs: Values to inject into loaded config.
    :return PredictiveTask[TaskPromptSignature, TaskResult, _TaskBridge]: Deserialized PredictiveTask instance.
    """
    init_dict = config.to_init_dict(cls, **kwargs)
    init_dict["model_settings"] = ModelSettings.model_validate(init_dict["model_settings"])

    return cls(**init_dict)

evaluate(docs, judge=None, failure_threshold=0.5)

Evaluate task performance using DSPy-based evaluation.

Parameters:

Name	Type	Description	Default
docs	Iterable[Doc]	Documents to evaluate.	required
judge	LM | None	Optional DSPy LM instance to use as judge for generative tasks.	None
failure_threshold	float	Decision threshold for whether to mark predicitions as failures.	0.5

Returns:

Type	Description
TaskEvaluationReport	Evaluation report.

Source code in sieves/tasks/predictive/core.py

def evaluate(
    self, docs: Iterable[Doc], judge: dspy.LM | None = None, failure_threshold: float = 0.5
) -> TaskEvaluationReport:
    """Evaluate task performance using DSPy-based evaluation.

    :param docs: Documents to evaluate.
    :param judge: Optional DSPy LM instance to use as judge for generative tasks.
    :param failure_threshold: Decision threshold for whether to mark predicitions as failures.
    :return: Evaluation report.
    """
    truths: list[Any] = []
    preds: list[Any] = []
    failures: list[Doc] = []

    # Evaluate each doc individually to identify failed predictions.
    for doc in docs:
        if self.id not in doc.results:
            continue

        pred = doc.results[self.id]
        gold = doc.gold.get(self.id, None)

        # Accumulate for corpus-level metrics.
        truths.append(gold)
        preds.append(pred)

        # If gold or prediction is None: we cannot do proper evalution, so we just check whether they're both None
        # to compute score for failure analysis.
        if gold is None or pred is None:
            if gold is not None or pred is not None:
                failures.append(doc)
        else:
            # Convert result and gold to DSPy representation.
            truth = dspy.Example(**self._task_result_to_dspy_dict(gold))
            pred_dspy = dspy.Prediction(**self._task_result_to_dspy_dict(pred))

            # Call internal evaluation logic for per-doc failure analysis.
            score = self._evaluate_dspy_example(truth, pred_dspy, trace=None, model=judge)

            if score < failure_threshold:
                failures.append(doc)

    # Evaluate on corpus level to obtain representative metrics.
    metrics = self._compute_metrics(truths, preds, judge=judge)

    return TaskEvaluationReport(
        metrics=metrics,
        task_id=self.id,
        failures=failures,
    )

optimize(optimizer, verbose=True)

Optimize task prompt and few-shot examples with the available optimization config.

Updates task to use best prompt and few-shot examples found by the optimizer.

Parameters:

Name	Type	Description	Default
optimizer	Optimizer	Optimizer to run.	required
verbose	bool	Whether to suppress output. DSPy produces a good amount of logs, so this can be useful to not pollute your terminal. Only warnings and errors will be printed.	True

Returns:

Type	Description
tuple[str, Sequence[FewshotExample]]	Best found prompt and few-shot examples.

Source code in sieves/tasks/predictive/core.py

def optimize(self, optimizer: optimization.Optimizer, verbose: bool = True) -> tuple[str, Sequence[FewshotExample]]:
    """Optimize task prompt and few-shot examples with the available optimization config.

    Updates task to use best prompt and few-shot examples found by the optimizer.

    :param optimizer: Optimizer to run.
    :param verbose: Whether to suppress output. DSPy produces a good amount of logs, so this can be useful to
        not pollute your terminal. Only warnings and errors will be printed.

    :return tuple[str, Sequence[FewshotExample]]: Best found prompt and few-shot examples.
    """
    assert len(self._fewshot_examples) > 1, "At least two few-shot examples need to be provided to optimize."

    # Run optimizer to get best prompt and few-shot examples.
    signature = self._get_task_signature()
    dspy_examples = [ex.to_dspy() for ex in self._fewshot_examples]

    def _pred_eval(truth: dspy.Example, pred: dspy.Prediction, trace: Any | None = None) -> float:
        """Wrap optimization evaluation, inject model.

        :param truth: Ground truth.
        :param pred: Predicted value.
        :param trace: Optional trace information.
        :return: Metric value between 0.0 and 1.0.
        :raises KeyError: If target fields are missing from truth or prediction.
        :raises ValueError: If similarity score cannot be parsed from LLM response.
        """
        return self._evaluate_dspy_example(truth, pred, trace, model=optimizer.model)

    if verbose:
        best_prompt, best_examples = optimizer(signature, dspy_examples, _pred_eval, verbose=verbose)
    else:
        # Temporarily suppress DSPy logs.
        dspy_logger = logging.getLogger("dspy")
        optuna_logger = logging.getLogger("optuna")
        original_dspy_level = dspy_logger.level
        original_optuna_level = optuna_logger.level

        try:
            dspy_logger.setLevel(logging.ERROR)
            optuna_logger.setLevel(logging.ERROR)
            with warnings.catch_warnings():
                warnings.simplefilter("ignore")
                best_prompt, best_examples = optimizer(signature, dspy_examples, _pred_eval, verbose=verbose)
        finally:
            dspy_logger.setLevel(original_dspy_level)
            optuna_logger.setLevel(original_optuna_level)

    # Update few-shot examples and prompt instructions.
    fewshot_example_cls = self._fewshot_examples[0].__class__
    self._fewshot_examples = [fewshot_example_cls.from_dspy(ex) for ex in best_examples]
    self._validate_fewshot_examples()
    self._custom_prompt_instructions = best_prompt

    # Reinitialize bridge to use new prompt and few-shot examples.
    self._bridge = self._init_bridge(ModelType.get_model_type(self._model_wrapper))

    return best_prompt, self._fewshot_examples

serialize()

Serialize task.

Returns:

Type	Description
Config	Config instance.

Source code in sieves/tasks/core.py

def serialize(self) -> Config:
    """Serialize task.

    :return: Config instance.
    """
    return Config.create(self.__class__, {k: Attribute(value=v) for k, v in self._state.items()})

Bridges for NER task.

DSPyNER

Bases: NERBridge[PromptSignature, Result, InferenceMode]

DSPy bridge for NER.

Source code in sieves/tasks/predictive/ner/bridges.py

class DSPyNER(NERBridge[dspy_.PromptSignature, dspy_.Result, dspy_.InferenceMode]):
    """DSPy bridge for NER."""

    @override
    def _validate(self) -> None:
        assert self._model_type == ModelType.dspy

    @override
    @property
    def model_type(self) -> ModelType:
        return ModelType.dspy

    @override
    @property
    def _default_prompt_instructions(self) -> str:
        return ""

    @override
    @property
    def inference_mode(self) -> dspy_.InferenceMode:
        return self._model_settings.inference_mode or dspy_.InferenceMode.predict

    @override
    def consolidate(
        self, results: Sequence[dspy_.Result], docs_offsets: list[tuple[int, int]]
    ) -> Sequence[dspy_.Result]:
        # Process each document (which may consist of multiple chunks)
        consolidated_results: list[dspy_.Result] = []
        for doc_offset in docs_offsets:
            doc_results = results[doc_offset[0] : doc_offset[1]]

            # Combine all entities from all chunks
            all_entities: list[Entity] = []

            # Process each chunk for this document
            for chunk_result in doc_results:
                if chunk_result is None:
                    continue

                if not hasattr(chunk_result, "entities") or not chunk_result.entities:
                    continue

                # Process entities in this chunk
                for entity in chunk_result.entities:
                    all_entities.append(entity)

            # Create a consolidated result for this document
            consolidated_results.append(
                dspy.Prediction.from_completions({"entities": [all_entities]}, signature=self.prompt_signature)
            )
        return consolidated_results

model_settings property

Return model settings.

Returns:

Type	Description
ModelSettings	Model settings.

prompt_template property

Return prompt template.

Chains _prompt_instructions, _prompt_example_xml and _prompt_conclusion.

Note: different model have different expectations as to how a prompt should look like. E.g. outlines supports the Jinja 2 templating format for insertion of values and few-shot examples, whereas DSPy integrates these things in a different value in the workflow and hence expects the prompt not to include these things. Mind model-specific expectations when creating a prompt template.

Returns:

Type	Description
str	Prompt template as string. None if not used by model wrapper.

__init__(task_id, prompt_instructions, entities, model_settings, prompt_signature, model_type, fewshot_examples=())

Initialize NER bridge.

Parameters:

Name	Type	Description	Default
task_id	str	Task ID.	required
prompt_instructions	str | None	Custom prompt instructions. If None, default instructions are used.	required
entities	list[str] | dict[str, str]	List of entities to extract or dict mapping labels to descriptions.	required
model_settings	ModelSettings	Settings for structured generation.	required
prompt_signature	type[BaseModel]	Unified Pydantic prompt signature.	required
model_type	ModelType	Model type.	required
fewshot_examples	Sequence[BaseModel]	Few-shot examples.	()

Source code in sieves/tasks/predictive/ner/bridges.py

def __init__(
    self,
    task_id: str,
    prompt_instructions: str | None,
    entities: list[str] | dict[str, str],
    model_settings: ModelSettings,
    prompt_signature: type[pydantic.BaseModel],
    model_type: ModelType,
    fewshot_examples: Sequence[pydantic.BaseModel] = (),
):
    """Initialize NER bridge.

    :param task_id: Task ID.
    :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
    :param entities: List of entities to extract or dict mapping labels to descriptions.
    :param model_settings: Settings for structured generation.
    :param prompt_signature: Unified Pydantic prompt signature.
    :param model_type: Model type.
    :param fewshot_examples: Few-shot examples.
    """
    assert model_type in {ModelType.dspy, ModelType.outlines, ModelType.langchain, ModelType.gliner}

    super().__init__(
        task_id=task_id,
        prompt_instructions=prompt_instructions,
        overwrite=False,
        model_settings=model_settings,
        prompt_signature=prompt_signature,
        model_type=model_type,
        fewshot_examples=fewshot_examples,
    )
    if isinstance(entities, dict):
        self._entities = list(entities.keys())
        self._entity_descriptions = entities
    else:
        self._entities = list(entities)
        self._entity_descriptions = {}

extract(docs)

Extract all values from doc instances that are to be injected into the prompts.

Parameters:

Name	Type	Description	Default
docs	Sequence[Doc]	Docs to extract values from.	required

Returns:

Type	Description
Sequence[dict[str, Any]]	All values from doc instances that are to be injected into the prompts as a sequence.

Source code in sieves/tasks/predictive/bridges.py

def extract(self, docs: Sequence[Doc]) -> Sequence[dict[str, Any]]:
    """Extract all values from doc instances that are to be injected into the prompts.

    :param docs: Docs to extract values from.
    :return: All values from doc instances that are to be injected into the prompts as a sequence.
    """
    return [{"text": doc.text if doc.text else None} for doc in docs]

NERBridge

Bases: Bridge[_BridgePromptSignature, _BridgeResult, ModelWrapperInferenceMode], ABC

Abstract base class for NER bridges.

Source code in sieves/tasks/predictive/ner/bridges.py

class NERBridge(Bridge[_BridgePromptSignature, _BridgeResult, ModelWrapperInferenceMode], abc.ABC):
    """Abstract base class for NER bridges."""

    def __init__(
        self,
        task_id: str,
        prompt_instructions: str | None,
        entities: list[str] | dict[str, str],
        model_settings: ModelSettings,
        prompt_signature: type[pydantic.BaseModel],
        model_type: ModelType,
        fewshot_examples: Sequence[pydantic.BaseModel] = (),
    ):
        """Initialize NER bridge.

        :param task_id: Task ID.
        :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
        :param entities: List of entities to extract or dict mapping labels to descriptions.
        :param model_settings: Settings for structured generation.
        :param prompt_signature: Unified Pydantic prompt signature.
        :param model_type: Model type.
        :param fewshot_examples: Few-shot examples.
        """
        assert model_type in {ModelType.dspy, ModelType.outlines, ModelType.langchain, ModelType.gliner}

        super().__init__(
            task_id=task_id,
            prompt_instructions=prompt_instructions,
            overwrite=False,
            model_settings=model_settings,
            prompt_signature=prompt_signature,
            model_type=model_type,
            fewshot_examples=fewshot_examples,
        )
        if isinstance(entities, dict):
            self._entities = list(entities.keys())
            self._entity_descriptions = entities
        else:
            self._entities = list(entities)
            self._entity_descriptions = {}

    @override
    @property
    def prompt_signature(self) -> _BridgePromptSignature:
        return convert_to_signature(
            model_cls=self._pydantic_signature,
            model_type=self.model_type,
            mode="entities",
        )  # type: ignore[return-value]

    def _get_entity_descriptions(self) -> str:
        """Return a string with the entity descriptions.

        :return: A string with the entity descriptions.
        """
        entities_with_descriptions: list[str] = []
        for entity in self._entities:
            if entity in self._entity_descriptions:
                entities_with_descriptions.append(
                    f"  <entity_description>\n    <entity>{entity}</entity>\n    <description>"
                    f"{self._entity_descriptions[entity]}</description>\n  </entity_description>"
                )
            else:
                entities_with_descriptions.append(f"  <entity>{entity}</entity>")

        entity_desc_string = "\n".join(entities_with_descriptions)
        return f"<entity_descriptions>\n{entity_desc_string}\n</entity_descriptions>"

    @staticmethod
    def _find_entity_positions(
        doc_text: str,
        result: _BridgeResult,
    ) -> list[Entity]:
        """Find all positions of an entity in a document.

        :param doc_text: The text of the document.
        :param result: The result of the model.
        :return: The list of entities with start/end indices.
        """
        doc_text_lower = doc_text.lower()
        # Create a new result with the same structure as the original
        new_entities: list[Entity] = []

        # Track entities by position to avoid duplicates
        entities_by_position: dict[tuple[int, int], Entity] = {}
        context_list: list[str] = []

        entities_list = getattr(result, "entities", [])
        for entity_with_context in entities_list:
            # Skip if there is no entity
            if not entity_with_context:
                continue

            # Get the entity and context texts from the model
            entity_text = getattr(entity_with_context, "text", "")
            context = getattr(entity_with_context, "context", None)
            entity_type = getattr(entity_with_context, "entity_type", "")
            score = getattr(entity_with_context, "score", None)

            if not entity_text:
                continue

            if context is None:
                new_entities.append(
                    Entity(
                        text=entity_text,
                        start=-1,
                        end=-1,
                        entity_type=entity_type,
                        score=score,
                    )
                )
                continue

            entity_text_lower = entity_text.lower()
            context_lower = context.lower() if context else ""
            # Create a list of the unique contexts
            # Avoid adding duplicates as entities witht he same context would be captured twice
            if context_lower not in context_list:
                context_list.append(context_lower)
            else:
                continue
            # Find all occurrences of the context in the document using regex
            context_positions = re.finditer(re.escape(context_lower), doc_text_lower)

            # For each context position that was found (usually is just one), find the entity within that context
            for match in context_positions:
                context_start = match.start()
                entity_start_in_context = context_lower.find(entity_text_lower)

                if entity_start_in_context >= 0:
                    start = context_start + entity_start_in_context
                    end = start + len(entity_text)

                    # Create a new entity with start/end indices
                    new_entity = Entity(
                        text=doc_text[start:end],
                        start=start,
                        end=end,
                        entity_type=entity_type,
                        score=score,
                    )

                    # Only add if this exact position hasn't been filled yet
                    position_key = (start, end)
                    if position_key not in entities_by_position:
                        entities_by_position[position_key] = new_entity
                        new_entities.append(new_entity)

        return sorted(new_entities, key=lambda x: x.start)

    @override
    def integrate(self, results: Sequence[_BridgeResult], docs: list[Doc]) -> list[Doc]:
        docs_list = list(docs)
        results_list = list(results)

        for doc, result in zip(docs_list, results_list):
            # Get the original text from the document
            doc_text = doc.text or ""
            if hasattr(result, "entities"):
                # Process entities from result if available
                entities_with_position = self._find_entity_positions(doc_text, result)
                # Create a new result with the updated entities
                new_result = Result(text=doc_text, entities=entities_with_position)
                doc.results[self._task_id] = new_result
            else:
                # Default empty result
                doc.results[self._task_id] = Result(text=doc_text, entities=[])

        return docs_list

inference_mode abstractmethod property

Return inference mode.

Returns:

Type	Description
ModelWrapperInferenceMode	Inference mode.

model_settings property

Return model settings.

Returns:

Type	Description
ModelSettings	Model settings.

model_type property

Return model type.

Returns:

Type	Description
ModelType	Model type.

prompt_template property

Return prompt template.

Chains _prompt_instructions, _prompt_example_xml and _prompt_conclusion.

Note: different model have different expectations as to how a prompt should look like. E.g. outlines supports the Jinja 2 templating format for insertion of values and few-shot examples, whereas DSPy integrates these things in a different value in the workflow and hence expects the prompt not to include these things. Mind model-specific expectations when creating a prompt template.

Returns:

Type	Description
str	Prompt template as string. None if not used by model wrapper.

__init__(task_id, prompt_instructions, entities, model_settings, prompt_signature, model_type, fewshot_examples=())

Initialize NER bridge.

Parameters:

Name	Type	Description	Default
task_id	str	Task ID.	required
prompt_instructions	str | None	Custom prompt instructions. If None, default instructions are used.	required
entities	list[str] | dict[str, str]	List of entities to extract or dict mapping labels to descriptions.	required
model_settings	ModelSettings	Settings for structured generation.	required
prompt_signature	type[BaseModel]	Unified Pydantic prompt signature.	required
model_type	ModelType	Model type.	required
fewshot_examples	Sequence[BaseModel]	Few-shot examples.	()

Source code in sieves/tasks/predictive/ner/bridges.py

def __init__(
    self,
    task_id: str,
    prompt_instructions: str | None,
    entities: list[str] | dict[str, str],
    model_settings: ModelSettings,
    prompt_signature: type[pydantic.BaseModel],
    model_type: ModelType,
    fewshot_examples: Sequence[pydantic.BaseModel] = (),
):
    """Initialize NER bridge.

    :param task_id: Task ID.
    :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
    :param entities: List of entities to extract or dict mapping labels to descriptions.
    :param model_settings: Settings for structured generation.
    :param prompt_signature: Unified Pydantic prompt signature.
    :param model_type: Model type.
    :param fewshot_examples: Few-shot examples.
    """
    assert model_type in {ModelType.dspy, ModelType.outlines, ModelType.langchain, ModelType.gliner}

    super().__init__(
        task_id=task_id,
        prompt_instructions=prompt_instructions,
        overwrite=False,
        model_settings=model_settings,
        prompt_signature=prompt_signature,
        model_type=model_type,
        fewshot_examples=fewshot_examples,
    )
    if isinstance(entities, dict):
        self._entities = list(entities.keys())
        self._entity_descriptions = entities
    else:
        self._entities = list(entities)
        self._entity_descriptions = {}

consolidate(results, docs_offsets) abstractmethod

Consolidate results for document chunks into document results.

Parameters:

Name	Type	Description	Default
results	Sequence[TaskResult]	Results per document chunk.	required
docs_offsets	list[tuple[int, int]]	Chunk offsets per document. Chunks per document can be obtained with results[docs_chunk_offsets[i][0]:docs_chunk_offsets[i][1]].	required

Returns:

Type	Description
Sequence[TaskResult]	Results per document as a sequence.

Source code in sieves/tasks/predictive/bridges.py

@abc.abstractmethod
def consolidate(self, results: Sequence[TaskResult], docs_offsets: list[tuple[int, int]]) -> Sequence[TaskResult]:
    """Consolidate results for document chunks into document results.

    :param results: Results per document chunk.
    :param docs_offsets: Chunk offsets per document. Chunks per document can be obtained with
        `results[docs_chunk_offsets[i][0]:docs_chunk_offsets[i][1]]`.
    :return: Results per document as a sequence.
    """

extract(docs)

Extract all values from doc instances that are to be injected into the prompts.

Parameters:

Name	Type	Description	Default
docs	Sequence[Doc]	Docs to extract values from.	required

Returns:

Type	Description
Sequence[dict[str, Any]]	All values from doc instances that are to be injected into the prompts as a sequence.

Source code in sieves/tasks/predictive/bridges.py

def extract(self, docs: Sequence[Doc]) -> Sequence[dict[str, Any]]:
    """Extract all values from doc instances that are to be injected into the prompts.

    :param docs: Docs to extract values from.
    :return: All values from doc instances that are to be injected into the prompts as a sequence.
    """
    return [{"text": doc.text if doc.text else None} for doc in docs]

PydanticNER

Bases: NERBridge[BaseModel, BaseModel, ModelWrapperInferenceMode], ABC

Base class for Pydantic-based NER bridges.

Source code in sieves/tasks/predictive/ner/bridges.py

class PydanticNER(NERBridge[pydantic.BaseModel, pydantic.BaseModel, ModelWrapperInferenceMode], abc.ABC):
    """Base class for Pydantic-based NER bridges."""

    @override
    def _validate(self) -> None:
        assert self._model_type in {ModelType.langchain, ModelType.outlines}

    @override
    @property
    def _default_prompt_instructions(self) -> str:
        entity_info = self._get_entity_descriptions() if self._entity_descriptions else ""
        return (
            "Your goal is to extract named entities from the text. Only extract entities of the specified types:\n"
            f"{self._entities}.\n"
            f"{entity_info}\n\n"
            "For each entity:\n"
            "- Extract the exact text of the entity\n"
            "- Include a SHORT context string that contains ONLY the entity and AT MOST 3 words before and 3 words "
            "after it.\n"
            "  DO NOT include the entire text as context. DO NOT include words that are not present in the original "
            "text\n"
            "  as introductory words (Eg. 'Text:' before context string).\n"
            "- Specify which type of entity it is (must be one of the provided entity types)\n"
            "- Provide a confidence score between 0.0 and 1.0 for the extraction.\n\n"
            "IMPORTANT:\n"
            "- If the same entity appears multiple times in the text, extract each occurrence separately with its own "
            "context"
        )

    @override
    @property
    def _prompt_conclusion(self) -> str | None:
        return "===========\n\n<text>{{ text }}</text>\n<entity_types>{{ entity_types }}</entity_types>\n<entities>"

    @override
    def consolidate(
        self, results: Sequence[pydantic.BaseModel], docs_offsets: list[tuple[int, int]]
    ) -> Sequence[pydantic.BaseModel]:
        assert issubclass(self.prompt_signature, pydantic.BaseModel)

        # Process each document (which may consist of multiple chunks).
        consolidated_results: list[pydantic.BaseModel] = []
        for doc_offset in docs_offsets:
            doc_results = results[doc_offset[0] : doc_offset[1]]

            # Combine all entities from all chunks
            all_entities: list[dict[str, Any]] = []

            # Process each chunk for this document
            for chunk_result in doc_results:
                if chunk_result is None:
                    continue

                if not hasattr(chunk_result, "entities") or not chunk_result.entities:
                    continue

                # Process entities in this chunk
                for entity in chunk_result.entities:
                    # We just need to combine all entities from all chunks
                    all_entities.append(entity)

            # Create a consolidated result for this document - instantiate the class with entities
            consolidated_results.append(self.prompt_signature(entities=all_entities))

        return consolidated_results

    @override
    @property
    def model_type(self) -> ModelType:
        return self._model_type

    @override
    @property
    def inference_mode(self) -> outlines_.InferenceMode | langchain_.InferenceMode:
        if self._model_type == ModelType.outlines:
            return self._model_settings.inference_mode or outlines_.InferenceMode.json
        elif self._model_type == ModelType.langchain:
            return self._model_settings.inference_mode or langchain_.InferenceMode.structured

        raise ValueError(f"Unsupported model type: {self._model_type}")

model_settings property

Return model settings.

Returns:

Type	Description
ModelSettings	Model settings.

prompt_template property

Return prompt template.

Chains _prompt_instructions, _prompt_example_xml and _prompt_conclusion.

Note: different model have different expectations as to how a prompt should look like. E.g. outlines supports the Jinja 2 templating format for insertion of values and few-shot examples, whereas DSPy integrates these things in a different value in the workflow and hence expects the prompt not to include these things. Mind model-specific expectations when creating a prompt template.

Returns:

Type	Description
str	Prompt template as string. None if not used by model wrapper.

__init__(task_id, prompt_instructions, entities, model_settings, prompt_signature, model_type, fewshot_examples=())

Initialize NER bridge.

Parameters:

Name	Type	Description	Default
task_id	str	Task ID.	required
prompt_instructions	str | None	Custom prompt instructions. If None, default instructions are used.	required
entities	list[str] | dict[str, str]	List of entities to extract or dict mapping labels to descriptions.	required
model_settings	ModelSettings	Settings for structured generation.	required
prompt_signature	type[BaseModel]	Unified Pydantic prompt signature.	required
model_type	ModelType	Model type.	required
fewshot_examples	Sequence[BaseModel]	Few-shot examples.	()

Source code in sieves/tasks/predictive/ner/bridges.py

def __init__(
    self,
    task_id: str,
    prompt_instructions: str | None,
    entities: list[str] | dict[str, str],
    model_settings: ModelSettings,
    prompt_signature: type[pydantic.BaseModel],
    model_type: ModelType,
    fewshot_examples: Sequence[pydantic.BaseModel] = (),
):
    """Initialize NER bridge.

    :param task_id: Task ID.
    :param prompt_instructions: Custom prompt instructions. If None, default instructions are used.
    :param entities: List of entities to extract or dict mapping labels to descriptions.
    :param model_settings: Settings for structured generation.
    :param prompt_signature: Unified Pydantic prompt signature.
    :param model_type: Model type.
    :param fewshot_examples: Few-shot examples.
    """
    assert model_type in {ModelType.dspy, ModelType.outlines, ModelType.langchain, ModelType.gliner}

    super().__init__(
        task_id=task_id,
        prompt_instructions=prompt_instructions,
        overwrite=False,
        model_settings=model_settings,
        prompt_signature=prompt_signature,
        model_type=model_type,
        fewshot_examples=fewshot_examples,
    )
    if isinstance(entities, dict):
        self._entities = list(entities.keys())
        self._entity_descriptions = entities
    else:
        self._entities = list(entities)
        self._entity_descriptions = {}

extract(docs)

Extract all values from doc instances that are to be injected into the prompts.

Parameters:

Name	Type	Description	Default
docs	Sequence[Doc]	Docs to extract values from.	required

Returns:

Type	Description
Sequence[dict[str, Any]]	All values from doc instances that are to be injected into the prompts as a sequence.

Source code in sieves/tasks/predictive/bridges.py

def extract(self, docs: Sequence[Doc]) -> Sequence[dict[str, Any]]:
    """Extract all values from doc instances that are to be injected into the prompts.

    :param docs: Docs to extract values from.
    :return: All values from doc instances that are to be injected into the prompts as a sequence.
    """
    return [{"text": doc.text if doc.text else None} for doc in docs]

Schemas for NER task.

Entities

Bases: BaseModel

Collection of entities with associated text.

Attributes: entities: List of entities. text: Source text.

Source code in sieves/tasks/predictive/schemas/ner.py

class Entities(pydantic.BaseModel):
    """Collection of entities with associated text.

    Attributes:
        entities: List of entities.
        text: Source text.
    """

    entities: list[Entity]
    text: str

Entity

Bases: BaseModel

Class for storing entity information.

Attributes: text: Entity text. start: Start offset. end: End offset. entity_type: Type of entity. score: Confidence score.

Source code in sieves/tasks/predictive/schemas/ner.py

class Entity(pydantic.BaseModel):
    """Class for storing entity information.

    Attributes:
        text: Entity text.
        start: Start offset.
        end: End offset.
        entity_type: Type of entity.
        score: Confidence score.
    """

    text: str
    start: int
    end: int
    entity_type: str
    score: float | None = None

    def __eq__(self, other: object) -> bool:
        """Compare two entities.

        :param other: Other entity to compare with.
        :return: True if entities are equal, False otherwise.
        """
        if not isinstance(other, Entity):
            return False
        return (
            self.start == other.start
            and self.end == other.end
            and self.text == other.text
            and self.entity_type == other.entity_type
        )

    def __hash__(self) -> int:
        """Compute entity hash.

        :returns: Entity hash.
        """
        return hash((self.start, self.end, self.text, self.entity_type))

__eq__(other)

Compare two entities.

Parameters:

Name	Type	Description	Default
other	object	Other entity to compare with.	required

Returns:

Type	Description
bool	True if entities are equal, False otherwise.

Source code in sieves/tasks/predictive/schemas/ner.py

def __eq__(self, other: object) -> bool:
    """Compare two entities.

    :param other: Other entity to compare with.
    :return: True if entities are equal, False otherwise.
    """
    if not isinstance(other, Entity):
        return False
    return (
        self.start == other.start
        and self.end == other.end
        and self.text == other.text
        and self.entity_type == other.entity_type
    )

__hash__()

Compute entity hash.

Returns:

Type	Description
int	Entity hash.

Source code in sieves/tasks/predictive/schemas/ner.py

def __hash__(self) -> int:
    """Compute entity hash.

    :returns: Entity hash.
    """
    return hash((self.start, self.end, self.text, self.entity_type))

EntityWithContext

Bases: BaseModel

Entity mention with its text span, context, and type.

Attributes: text: The specific text segment identified as an entity. context: The surrounding text providing context for the entity. entity_type: The category or type of the entity. score: Confidence score for the entity identification.

Source code in sieves/tasks/predictive/schemas/ner.py

class EntityWithContext(pydantic.BaseModel):
    """Entity mention with its text span, context, and type.

    Attributes:
        text: The specific text segment identified as an entity.
        context: The surrounding text providing context for the entity.
        entity_type: The category or type of the entity.
        score: Confidence score for the entity identification.
    """

    text: str = pydantic.Field(description="The specific text segment identified as an entity.")
    context: str = pydantic.Field(description="The surrounding text providing context for the entity.")
    entity_type: str = pydantic.Field(description="The category or type of the entity (e.g., PERSON, ORGANIZATION).")
    score: float | None = pydantic.Field(
        default=None, description="Provide a confidence score for the entity identification, between 0 and 1."
    )

FewshotExample

Bases: FewshotExample

Few‑shot example with entities annotated in text.

Attributes: text: Input text. entities: List of entities with context.

Source code in sieves/tasks/predictive/schemas/ner.py

class FewshotExample(BaseFewshotExample):
    """Few‑shot example with entities annotated in text.

    Attributes:
        text: Input text.
        entities: List of entities with context.
    """

    text: str
    entities: list[EntityWithContext]

    @property
    def target_fields(self) -> tuple[str, ...]:
        """Return target fields.

        :return: Target fields.
        """
        return ("entities",)

input_fields property

Defines which fields are inputs.

Returns:

Type	Description
Sequence[str]	Sequence of field names.

target_fields property

Return target fields.

Returns:

Type	Description
tuple[str, ...]	Target fields.

from_dspy(example) classmethod

Convert from dspy.Example.

Parameters:

Name	Type	Description	Default
example	Example	Example as dspy.Example.	required

Returns:

Type	Description
Self	Example as FewshotExample.

Source code in sieves/tasks/predictive/schemas/core.py

@classmethod
def from_dspy(cls, example: dspy.Example) -> Self:
    """Convert from `dspy.Example`.

    :param example: Example as `dspy.Example`.
    :returns: Example as `FewshotExample`.
    """
    return cls(**example)

to_dspy()

Convert to dspy.Example.

Returns:

Type	Description
Example	Example as dspy.Example.

Source code in sieves/tasks/predictive/schemas/core.py

def to_dspy(self) -> dspy.Example:
    """Convert to `dspy.Example`.

    :returns: Example as `dspy.Example`.
    """
    return dspy.Example(**ModelWrapper.convert_fewshot_examples([self])[0]).with_inputs(*self.input_fields)

Result

Bases: Entities

Result of a named-entity recognition (NER) task. Contains the extracted entities and the source text.

Source code in sieves/tasks/predictive/schemas/ner.py

class Result(Entities):
    """Result of a named-entity recognition (NER) task. Contains the extracted entities and the source text."""

    pass