PdfElixide.Document.StructuredPage.Region (pdf_elixide v0.23.0)

A group of spans sharing one role on a structured page, with their joined text and union bounding box.

kind/0 lists the roles. Spans are grouped by role, column and section. Body text and marginal labels can have columns: 0 for left and 1 for right, each read top to bottom.

Gaps, blank lines and indents do not split regions, so consecutive body paragraphs can merge. Labels in the same column can also merge across intervening body text. To read paragraphs, use PdfElixide.Document.to_plain_text/2, or split spans at wider gaps between their boxes.

Regions follow their first spans in top-to-bottom, then left-to-right order. A right column starting higher appears before the left column; group or sort by column when you need left-column-then-right reading order.

bbox is the union of the spans' boxes in the page's raw, unrotated user space, like PdfElixide.Document.spans/2 — see Rotated pages and extracted geometry. section is set only on a tagged page, from the nearest /Sect, /Art or /Part structure element, and is stable across pages so a chapter continuing onto the next page keeps one value.

Summary

Types

kind()

The role of a region.

t()

Types

kind()

@type kind() ::
  :body
  | :marginal_label
  | :header
  | :footer
  | :page_number
  | :artifact
  | {:heading, 1..6}

The role of a region.

:body — ordinary text, including any margin text the document does not mark as an artifact.
:marginal_label — a short standalone numeral such as a verse or section number: either text of at most four digits or lowercase roman letters, or content a tagged PDF marks /Lbl.
:header, :footer, :page_number, :artifact — content the document marks as an /Artifact with /Type /Pagination and a /Subtype of /Header, /Footer or /PageNumber; any other artifact is :artifact. These roles come from that marking alone: a running header that merely sits at the top of the page is :body.
{:heading, level} — a tagged /H1 to /H6 heading, levels 1 to 6. This library does not produce it: such a heading arrives as :body.

t()

@type t() :: %PdfElixide.Document.StructuredPage.Region{
  bbox: PdfElixide.Geometry.Rect.t(),
  column: non_neg_integer() | nil,
  kind: kind(),
  page: non_neg_integer(),
  section: non_neg_integer() | nil,
  spans: [PdfElixide.Document.Span.t()],
  text: String.t()
}