Q:
what exactly is a “flow”? The description of TextExtractor::Line::GetFlowID
()
says:
The unique identifier for a paragraph or column that this line belongs to.
This is confusing as columns are usually made up of paragraphs.
Do you have a more precise definition of “flow”? For instance, does a flow
always have rectangular shape?
A:
Flow is a logical construct used to represent a reading order. A flow contains a sequences of blocks (paragraphs / columns) representing the reading order on a given page.
Most pages contain only a single flow, but it could contain 2 or more – e.g. if half of the page is portrait and the content on the other half is rotated 90 degrees.