bnl.core.MultiSegment

class bnl.core.MultiSegment(raw_layers=<factory>, *, name=None)

Bases: TimeSpan

The primary input object for analysis, containing multiple Segment layers.

Public Data Attributes:

itvls	Returns a list of all the intervals for each layer in the MultiSegment.
labels	Returns a list of all the labels for each layer in the MultiSegment.
raw_layers	A sequence of Segment objects representing different layers of annotation.
start
end
name

Inherited from TimeSpan

name
duration	Duration in seconds.
start
end

Public Methods:

layers	Returns the layers aligned to a unified time span.
from_json(json_data[, name])	Data Ingestion from adobe json format.
from_itvls(itvls, labels[, name])
plot([colorscale, hatch])	Plots the MultiSegment on a Plotly figure.
contour([strategy])	Calculates boundary salience and converts to a BoundaryContour.
scrub_labels()	Scrubs the labels of the MultiSegment by replacing them with empty strings.
find_span(layers[, mode])	Finds the span of a list of Segment layers.
align(span)	Align with a TimeSpan object.
prune_layers([relabel])	Prunes identical layers from the MultiSegment.
squeeze_layers([times, relabel])	Remove the least informative layer from the MultiSegment according to vmeasure.

---

Parameters:

• raw_layers (Sequence[Segment])

• name (str | None)

raw_layers: Sequence[Segment]

A sequence of Segment objects representing different layers of annotation.

property layers: Sequence[Segment]

Returns the layers aligned to a unified time span.

property itvls: Sequence[ndarray]

Returns a list of all the intervals for each layer in the MultiSegment.

property labels: Sequence[Sequence[str]]

Returns a list of all the labels for each layer in the MultiSegment.

classmethod from_json(json_data, name=None)

Data Ingestion from adobe json format.

Parameters:

• json_data (list) – A list of layers, where each layer is a tuple of (intervals, labels). intervals is a list of [start, end] times, and labels is a list of strings.

• name (str, optional) – Name for the created MultiSegment.

Return type:

MultiSegment

plot(colorscale='D3', hatch=True)

Plots the MultiSegment on a Plotly figure.

Parameters:

• colorscale (str | list[str], optional) – Plotly colorscale to use. Can be a qualitative scale name (e.g., “Set3”, “Pastel”) or a list of colors.

• hatch (bool, optional) – Whether to use hatch patterns for different labels. Defaults to True.

Return type:

Figure

contour(strategy='depth', **kwargs)

Calculates boundary salience and converts to a BoundaryContour.

Parameters:

• strategy (str)

• kwargs (Any)

Return type:

BoundaryContour

scrub_labels()

Scrubs the labels of the MultiSegment by replacing them with empty strings.

Return type:

MultiSegment

static find_span(layers, mode='common')

Finds the span of a list of Segment layers.

Parameters:

• mode (str, optional) – The alignment mode. Can be “union” or “common”. Defaults to “common”.

• layers (Sequence[Segment])

Return type:

TimeSpan

align(span)

Align with a TimeSpan object.

Parameters:

span (TimeSpan)

Return type:

MultiSegment

prune_layers(relabel=True)

Prunes identical layers from the MultiSegment.

This also gets rid of layers with no inner boundaries.

Parameters:

relabel (bool)

Return type:

MultiSegment

squeeze_layers(times=1, relabel=True)

Remove the least informative layer from the MultiSegment according to vmeasure.

Returns a new MultiSegment with the most redundant layer removed.

Parameters:

• times (int)

• relabel (bool)

Return type:

MultiSegment