simple-seq

One of two sequence-like formats that are meant to be general, and allow for working with a wide array of formats – the other is generic-seq.

The 'simple-seq' format can represent any sequence-like format,
that can be mapped to a sequence of segments, with each segment having an onset time, offset time, and label.

A file in 'simple-seq' format can be either a .csv or .txt file. It should have 3 columns that represent the onset and offset times in seconds and the labels of the segments in the annotated sequences.

The default is to assume a comma-separated values file with a header ‘onset_s, offset_s, label’, but this can be modified with keyword arguments to the from_file method of the class that represents the format.

This format also assumes that each annotation file corresponds to one annotated source file, i.e. a single audio or spectrogram file. This is likely to be the case if you are using apps like Praat or Audacity. An example of such a format is the Audacity standard label track format, exported to .txt files, that you would get if you were to annotate with region labels.

The annnotations can be loaded with the following class: crowsetta.formats.seq.simple.SimpleSeq.