PsychiatryNLPKit.analysis.AnalysisSpec

class PsychiatryNLPKit.analysis.AnalysisSpec(name, data_source, func=None, loader=None, build_kwargs=None, requires=(), manages_lifecycle=False, requires_extra=None)

Declarative description of a registered analysis.

Attributes:

name: Public analysis name (used by TextData.compute and

BatchAnalyzer).

data_source: Name of the TextData attribute providing the

function’s primary argument, "data" for the raw text dict, or None when the function receives everything as kwargs.

build_kwargs: Callable returning extra keyword arguments derived

from a TextData instance (language, model references).

requires: TextData model attributes the analysis needs loaded.

manages_lifecycle: Whether the function has a manage_lifecycle

parameter (model-backed analyses).

requires_extra: Name of the optional install extra the analysis

needs ("graph" or "image"), or None.

__init__(name, data_source, func=None, loader=None, build_kwargs=None, requires=(), manages_lifecycle=False, requires_extra=None)

Methods

__init__(name, data_source[, func, loader, ...])
resolve()	Return the underlying analysis callable (lazy for optional deps).

Attributes

build_kwargs
func
loader
manages_lifecycle
requires
requires_extra
name
data_source

name: str

data_source: str | None

func: Callable[[...], dict[str, dict[str, float]]] | None = None

loader: Callable[[], Callable[[...], dict[str, dict[str, float]]]] | None = None

build_kwargs: Callable[[Any], dict] | None = None

requires: tuple[str, ...] = ()

manages_lifecycle: bool = False

requires_extra: str | None = None

resolve()

Return the underlying analysis callable (lazy for optional deps).

__init__(name, data_source, func=None, loader=None, build_kwargs=None, requires=(), manages_lifecycle=False, requires_extra=None)