crowsetta.Sequence

class crowsetta.Sequence(segments, labels, onsets_s=None, offsets_s=None, onset_samples=None, offset_samples=None)

Bases: object

A class that represents a sequence of segments, used to annotate animal acoustic communication.

E.g., a human sentence made up of syllables, or a bout of birdsong made up of “syllables”.

segments

A tuple of crowsetta.Segment instances.

Type:

tuple

onset_samples

Numpy array of type int, onset of each annotated segment in sample number.

Type:

numpy.ndarray or None

offset_samples

Numpy array of type int, offset of each annotated segment in sample number.

Type:

numpy.ndarray or None

onsets_s

Numpy array of type float, onset of each annotated segment in seconds.

Type:

numpy.ndarray or None

offsets_s

Numpy array of type float, offset of each annotated segment in seconds.

Type:

numpy.ndarray or None

labels

Numpy array of type char, label for each annotated segment.

Type:

str, list, or numpy.ndarray

from_segments : method

Make a Sequence from a list of :class:`~crowsetta.Segment`s.

from_keyword : method

Make a Sequence by passing keywords (all arguments except segments)

from_dict : method

Like from_keyword, but pass a Python dictionary where keys are keywords and values are arguments for those keywords.

to_dict : method

Convert to a dict. The inverse of from_dict().

Examples

A sequence with onsets and offsets given in seconds.

>>> import numpy as np
>>> import crowsetta
>>> onsets_s = np.array([1.0, 3.0, 5.0])
>>> offsets_s = np.array(2.0, 4.0, 6.0])
>>> labels = np.array(['a', 'a', 'b'])
>>> seq = crowsetta.Sequence.from_keyword(labels=labels, onsets_s=onsets_s, offsets_s=offsets_s)

The same sequence could also be made by calling the from_segments() class method.

>>> segments = []
>>> for onset, offset, label in zip(onsets_s, offsets_s, labels):
...     segments.append(crowsetta.Segment(onset_s=onset, offset_s=offset, label=label))
>>> seq2 = crowsetta.Sequence.from_segments(segments)

__init__(segments, labels, onsets_s=None, offsets_s=None, onset_samples=None, offset_samples=None)

Initialize a new Sequence instance.

Parameters:

• segments (tuple) – A tuple of crowsetta.Segment instances.

• onset_samples (numpy.ndarray or None) – Numpy array of type int, onset of each annotated segment in sample number.

• offset_samples (numpy.ndarray or None) – Numpy array of type int, offset of each annotated segment in sample number.

• onsets_s (numpy.ndarray or None) – Numpy array of type float, onset of each annotated segment in seconds.

• offsets_s (numpy.ndarray or None) – Numpy array of type float, offset of each annotated segment in seconds.

• labels (str, list, or numpy.ndarray) – Numpy array of type char, label for each annotated segment.

Methods

__init__(segments, labels[, onsets_s, ...])	Initialize a new Sequence instance.
as_dict()	Convert this crowsetta.Sequence to a dict.
from_dict(seq_dict)	Construct a crowsetta.Sequence from a dict where keys are arguments to from_keyword().
from_keyword(labels[, onset_samples, ...])	Construct a crowsetta.Sequence from keyword arguments
from_segments(segments)	Construct a crowsetta.Sequence from a list of crowsetta.Segment objects.

Attributes

labels
offset_samples
offsets_s
onset_samples
onsets_s
segments

as_dict() → dict

Convert this crowsetta.Sequence to a dict.

Returns:

seq_dict –

with the following key, value pairs:

onset_samplesnumpy.ndarray or None

of type int, onset of each annotated segment in samples/second

offset_samplesnumpy.ndarray or None

of type int, offset of each annotated segment in samples/second

onsets_snumpy.ndarray or None

of type float, onset of each annotated segment in seconds

offsets_snumpy.ndarray or None

of type float, offset of each annotated segment in seconds

labelsnumpy.ndarray

of type str; label for each annotated segment

Return type:

dict

classmethod from_dict(seq_dict)

Construct a crowsetta.Sequence from a dict where keys are arguments to from_keyword().

Parameters:

• seq_dict (dict) –

  with following key, value pairs onset_samples : numpy.ndarray or None

  of type int, onset of each annotated segment in samples/second

  offset_samplesnumpy.ndarray or None

  of type int, offset of each annotated segment in samples/second

  onsets_snumpy.ndarray or None

  of type float, onset of each annotated segment in seconds

  offsets_snumpy.ndarray or None

  of type float, offset of each annotated segment in seconds

  labelsstr, list, or numpy.ndarray

  of type str, label for each annotated segment

• offsets (seq_dict must specify both onsets and) –

:param : :param either in units of samples or seconds (or both).:

Examples

>>> seq_dict = {
...     'labels': 'abc',
...     'onset_samples': np.asarray([16005, 17925, 19837]),
...     'offset_samples': np.asarray([17602, 19520, 21435]),
...     'file': 'bird0.wav',
...     }
>>> seq = Sequence.from_dict(seq_dict)

classmethod from_keyword(labels, onset_samples=None, offset_samples=None, onsets_s=None, offsets_s=None)

Construct a crowsetta.Sequence from keyword arguments

Parameters:

• onset_samples (numpy.ndarray or None) – of type int, onset of each annotated segment in samples/second

• offset_samples (numpy.ndarray or None) – of type int, offset of each annotated segment in samples/second

• onsets_s (numpy.ndarray or None) – of type float, onset of each annotated segment in seconds

• offsets_s (numpy.ndarray or None) – of type float, offset of each annotated segment in seconds

• labels (str, list, or numpy.ndarray) – of type str, label for each annotated segment

• offsets (Must specify both onsets and) –

:param : :param either in units of Hz or seconds (or both).:

classmethod from_segments(segments)

Construct a crowsetta.Sequence from a list of crowsetta.Segment objects.

Parameters:

segments (list) – A list of crowsetta.Segment instances.

Returns:

seq – A Sequence instance generated using the list of :class:`~crowsetta.Segment`s.

Return type:

crowsetta.Sequence