how to use transcriptions from .cha

[alecristia]

.cha transcriptions often include non-verbal material, such as @l to indicate that something is pronounced as a letter, or @spa to tag a single word as Spanish. These marks are there even in the converted versions

How do we clean up such transcriptions?

[alecristia]

there are at least two ways:

• if you want to clean up the transcriptions to leave only the lexical content, then use the pylangacq package -- notice that you have a couple of options RE how to extract words, or lexemes, etc
• if you want to clean up to get the phonological form instead, then it may be better to look at what was done in wordseg: https://docs.cognitive-ml.fr/wordseg/ -- for instance, this project built on .cha files