vertebrae.adapters

Typed structured-unit adapters built on top of UnitAnnotation workflows.

Classes

`RegionAnnotation`	Per-parent detection or document-layout region annotations.
`SequenceAnnotation`	Per-parent OCR, ASR, or sequence-labeling annotations.
`KeypointAnnotation`	Per-parent pose or keypoint annotations.
`DepthAnnotation`	Per-parent sampled depth annotations for regression-style unit targets.
`LatentSlotAnnotation`	Per-parent latent-slot or generative structured-unit annotations.
`DetectionLayoutAdapter`	Normalize detection and layout annotations into structured unit annotations.
`SequenceLabelingAdapter`	Normalize OCR, ASR, and token-sequence annotations.
`KeypointAdapter`	Normalize keypoint annotations into structured unit annotations.
`DepthAdapter`	Normalize sampled depth annotations into regression-style structured units.
`LatentSlotAdapter`	Normalize latent-slot annotations into structured unit annotations.

class vertebrae.adapters.RegionAnnotation[source]: Per-parent detection or document-layout region annotations.

class vertebrae.adapters.SequenceAnnotation[source]: Per-parent OCR, ASR, or sequence-labeling annotations.

class vertebrae.adapters.KeypointAnnotation[source]: Per-parent pose or keypoint annotations.

class vertebrae.adapters.DepthAnnotation[source]: Per-parent sampled depth annotations for regression-style unit targets.

class vertebrae.adapters.LatentSlotAnnotation[source]: Per-parent latent-slot or generative structured-unit annotations.

class vertebrae.adapters.DetectionLayoutAdapter(unit_type='region', task_family='detection_layout')[source]

Normalize detection and layout annotations into structured unit annotations.

Parameters:

class vertebrae.adapters.SequenceLabelingAdapter(unit_type='token', task_family='sequence')[source]

Normalize OCR, ASR, and token-sequence annotations.

Parameters:

class vertebrae.adapters.KeypointAdapter(unit_type='keypoint', task_family='keypoint')[source]

Normalize keypoint annotations into structured unit annotations.

Parameters:

class vertebrae.adapters.DepthAdapter(unit_type='depth_sample', task_family='depth')[source]

Normalize sampled depth annotations into regression-style structured units.

Parameters:

class vertebrae.adapters.LatentSlotAdapter(unit_type='latent_slot', task_family='latent_slot')[source]

Normalize latent-slot annotations into structured unit annotations.

Parameters: