Aligned TextGrid

Docs Last Built

2024-11-13 16:10:56+00:00

PyPI Build codecov Maintainability Build Docs DOI

The aligned-textgrid package provides a python interface for representing and operating on TextGrids produced by forced aligners like FAVE or the Montreal Forced Aligner. Classes provided by aligned-textgrid represent hierarchical and precedence relationships among data stored in TextGrid formats allowing for simplified and more accessible analysis of aligned speech data.

Example Use Cases

  • You want to quickly loop through the Phone tier of a Textgrid, and also access information about the word it is a part of.
  • You want to quickly loop over the Word tier of a Textgrid and quickly count how many phones it has.
  • You want to programmatically merge together adjacent Textgrid intervals.

For examples on how to use the pacakge, see the Usage pages.

Installation

To install aligned-textgrid using pip, run the following command in your terminal:

pip install aligned-textgrid

Not another TextGrid implementation

There are several other packages that parse Praat Textgrids, including

aligned-textgrid’s goal is to capture hierarchical and sequential relationships represented in many TextGrids, and to make them easilly accessible to users via an intuitive interface. The goal is that from any arbitrary location within a TextGrid, users can easilly access information with minimally defensive coding.

Example

As an example, we’ll read in a textgrid produced with forced alignment that contains a single speaker with a word and phone tier.

from aligned_textgrid import AlignedTextGrid, Word, Phone
tg = AlignedTextGrid(
    textgrid_path='usage/resources/josef-fruehwald_speaker.TextGrid', 
    entry_classes=[Word, Phone]
    )

Then, we can access an arbitrary phone interval.

arbitrary_interval = tg[0].Phone[20]

From this aribitrary interval, we can access information about the intervals preceding and following with the .prev and .fol attributes.

print(arbitrary_interval.prev.label)
print(arbitrary_interval.label)
print(arbitrary_interval.fol.label)
R
EY1
N

We can also access information about the word this interval is nested within with the .inword attribute.

print(arbitrary_interval.inword.label)
raindrops

The object returned by .inword is just another interval, meaning we can access informaton about it’s context with the .prev and .fol attributes as well.

print(arbitrary_interval.inword.prev.label)
print(arbitrary_interval.inword.label)
print(arbitrary_interval.inword.fol.label)
strikes
raindrops
in

For more

You can also directly read up on the function and class references.

Reuse

GPLv3