pagemark

Overview ¶

Package pagemark is the one place that knows how the pages of a document are marked in the markdown that goes to the chunkers.

A PDF is read page by page, and a chunk must be able to say which pages it comes from (for a citation, or to open the page). The text that is indexed therefore has, before the text of each page, a line that an HTML comment makes invisible to anyone who renders the markdown:

<!-- page 3 -->

The chunkers read these lines, take them out of the text, and put the pages a chunk covers in its metadata.