mava_exchange.tracks.AnnotationListSeries¶
- class mava_exchange.tracks.AnnotationListSeries(name: str, description: str, parent: str | None = None, derived_from: list[str] | None = None, method: str | None = None)¶
Sparse interval annotations with multiple labels per segment.
Use for multi-label scene tags, keywords, or any annotation where multiple values apply simultaneously to a time segment.
Each row in the Parquet file needs: start_seconds, end_seconds, annotations (as Python lists).
Example:
scene_tags = AnnotationListSeries( name="scene_tags", description="Scene classification (indoor/outdoor + natural/man-made)" ) df = pd.DataFrame({ "start_seconds": [0.0, 45.2, 78.5], "end_seconds": [45.2, 78.5, 120.0], "annotations": [ ["outdoor", "natural"], ["indoor"], ["outdoor", "man-made"] ] })
- __init__(name: str, description: str, parent: str | None = None, derived_from: list[str] | None = None, method: str | None = None) None¶
Methods
Attributes
Returns column names.
Source track names this track is derived from (provenance).
Derivation method (e.g. 'argmax', 'cluster_to_scalar', 'aggregate_scalar').
Containment parent track name (None if top-level).
Track name (must be unique within package).
Human-readable description of what this track contains.
- name: str¶
Track name (must be unique within package).
- description: str¶
Human-readable description of what this track contains.
- parent: str | None = None¶
Containment parent track name (None if top-level).
- derived_from: list[str] | None = None¶
Source track names this track is derived from (provenance).
- method: str | None = None¶
Derivation method (e.g. ‘argmax’, ‘cluster_to_scalar’, ‘aggregate_scalar’).
- type: Literal['mava:AnnotationListSeries'] = 'mava:AnnotationListSeries'¶
- property columns: list[str]¶
Returns column names.
- to_dict() dict¶
Converts to dictionary for manifest.json.