Reads a Word transcript. The extractor is handed a copy of the file's own bytes (a view such as a Node Buffer may
share its memory with other data). The text it answers is read with each paragraph a line, by the caller's layout,
or, without one, by Otter's headers when it is laid out as Otter's export (a header and its words in paragraphs of
their own) and otherwise as plain lines, Name: words naming the speakers by the shared rule. A Word file with no
text reads as no turns. A Word file with no empty paragraphs holds no blank line between its turns, so its first
header alone starts a block: a voice note's export there reads by the label each header gives, a later paragraph of a
turn that ends in a time as words, and a second speaker's header whose label the first header does not give as words
too, as readOtter sets out with the input.
Reads a Word transcript. The extractor is handed a copy of the file's own bytes (a view such as a Node
Buffermay share its memory with other data). The text it answers is read with each paragraph a line, by the caller'slayout, or, without one, by Otter's headers when it is laid out as Otter's export (a header and its words in paragraphs of their own) and otherwise as plain lines,Name: wordsnaming the speakers by the shared rule. A Word file with no text reads as no turns. A Word file with no empty paragraphs holds no blank line between its turns, so its first header alone starts a block: a voice note's export there reads by the label each header gives, a later paragraph of a turn that ends in a time as words, and a second speaker's header whose label the first header does not give as words too, as readOtter sets out with the input.