Function analyze_text

openmed/__init__.py:150–604

Run a token-classification model on ``text`` and format the predictions. Args: text: Clinical or biomedical text to analyse. model_name: Registry key, fully-qualified Hugging Face model id, or local model path. model_id: Alias for ``model_name``. Useful for A

(
    text: str,
    model_name: str = "disease_detection_superclinical",
    *,
    model_id: Optional[str] = None,
    config: Optional[OpenMedConfig] = None,
    loader: Optional[ModelLoader] = None,
    aggregation_strategy: Optional[str] = "simple",
    output_format: str = "dict",
    include_confidence: bool = True,
    confidence_threshold: Optional[float] = 0.0,
    group_entities: bool = False,
    formatter_kwargs: Optional[Dict[str, Any]] = None,
    metadata: Optional[Dict[str, Any]] = None,
    use_fast_tokenizer: bool = True,
    sentence_detection: bool = True,
    sentence_language: str = "en",
    sentence_clean: bool = False,
    sentence_segmenter: Optional[Any] = None,
    cache_results: bool = False,
    max_cache_entries: int = 128,
    **pipeline_kwargs: Any,
)

def analyze_text(
    text: str,
    model_name: str = "disease_detection_superclinical",
    *,
    model_id: Optional[str] = None,
    config: Optional[OpenMedConfig] = None,
    loader: Optional[ModelLoader] = None,
    aggregation_strategy: Optional[str] = "simple",
    output_format: str = "dict",
    include_confidence: bool = True,
    confidence_threshold: Optional[float] = 0.0,
    group_entities: bool = False,
    formatter_kwargs: Optional[Dict[str, Any]] = None,
    metadata: Optional[Dict[str, Any]] = None,
    use_fast_tokenizer: bool = True,
    sentence_detection: bool = True,
    sentence_language: str = "en",
    sentence_clean: bool = False,
    sentence_segmenter: Optional[Any] = None,
    cache_results: bool = False,
    max_cache_entries: int = 128,
    **pipeline_kwargs: Any,
) -> Union[AnalyzeResult, str, List[Dict[str, Any]]]:
    """Run a token-classification model on ``text`` and format the predictions.

    Args:
        text: Clinical or biomedical text to analyse.
        model_name: Registry key, fully-qualified Hugging Face model id, or
            local model path.
        model_id: Alias for ``model_name``. Useful for APIs and examples that
            name model identifiers as ``model_id``.
        config: Optional :class:`~openmed.core.config.OpenMedConfig` instance.
        loader: Reuse an existing :class:`~openmed.core.models.ModelLoader`.
        aggregation_strategy: Hugging Face aggregation strategy (``"simple"`` by
            default). Set to ``None`` to work with raw token outputs.
        output_format: ``"dict"`` (default), ``"json"``, ``"html"`` or ``"csv"``.
        include_confidence: Whether to include confidence scores in formatted output.
        confidence_threshold: Minimum confidence for entities. ``None`` keeps all.
        group_entities: Merge adjacent entities of the same label in the formatted
            output.
        formatter_kwargs: Extra keyword arguments forwarded to
            :func:`openmed.processing.format_predictions`.
        metadata: Optional metadata to attach to the result.
        use_fast_tokenizer: Prefer fast tokenizers when available.
        sentence_detection: Enable pySBD-powered sentence detection (default: True).
        sentence_language: Language hint for the sentence detector.
        sentence_clean: Whether to enable pySBD's cleaning heuristics.
        sentence_segmenter: Optional preconstructed pySBD segmenter to reuse.
        cache_results: Whether to cache this result in the in-process LRU cache. Cached results may contain PHI, but are never saved to disk.
        max_cache_entries: Maximum number of cached results.
        **pipeline_kwargs: Additional arguments passed to
            :meth:`openmed.core.models.ModelLoader.create_pipeline`.

    Returns:
        Analyze result for ``"dict"`` output, otherwise the requested rendered
        format.

    Example: