Add support for obtaining Transcript segments in sentences chunks.

Using the current system for obtaining Transcript segments, it is almost too difficult to obtain good context by setting the ```chunkSize```  to any value. It will be more intuitive for the segments to have an option to be based on a perceived sentence. 