ryantate/typingpool

Enable overlap

Opened this issue · 0 comments

Splitting by a consistent chunk length invariably leads to words cut out at beginning and end of transcripts.
Is it possible to code in an overlap function, maybe by percentage?

Example:
Transcript is 1 hour (60min).
Split into 60 x 1 min chunks or use the funk variable to calculate for the below.
Allow 5% (variable) overlap to give 63 second chunks, 1.5 secs each side.
Output 63 second chunks, centred on the 60 second markers from initial chunk calculation so give %variable overlap at right place.

Can it be done?