Crate pdfboss_output 

Layout analysis and output rendering for pdfboss: turns pdfboss-text spans into a structured layout IR, and the IR into a document.

Structs

BBox

A device-space box: y grows upward, as in PDF user space.

Cell

One table cell. An empty cell — or one covered by a neighbour’s span — carries no line.

ExtractReport

What extraction could not read. Extraction is lenient the way rendering is — content that will not fetch, decode, or parse yields no text rather than an error — and this report is what keeps that leniency accountable: an empty result with an empty report really is an empty page.

FontCache

Loaded fonts shared across page extractions of one document, keyed by the font dictionary’s object reference.

Inline

A run of same-styled text within a line. text already carries the spaces the word-gap rule inserted, so rendering a line is concatenation.

Line

One visual line. The geometry travels with it because later structure passes — lists, tables, page headers and footers — classify lines by it.

ListItem

One list item: its marker, the marker text’s length in characters (the continuation indent a wrapped item is measured against), and its lines.

Markdown

Markdown: ATX headings ranked by font size, one output line per source line, and emphasis around each run of styled text. Blocks — across pages too — are separated by a blank line.

PageLayout

One page’s blocks in reading order.

Ruling

An axis-aligned line segment a page draws, in the same y-up user space as TextSpan: a table border, a separator, an underline.

SkippedText

One piece of content whose text (if any) is missing from the result.

Text

Plain text: every line of every block in reading order joined with \n, pages separated by a form feed. Structure is invisible here — this is the adapter that must stay byte-equal to positional text extraction.

TextSpan

A positioned run of extracted text.

Enums

Block

One structural unit of a page, in reading order.

Marker

What introduces a list item.

Role

What a paragraph is to the page: its body, or a page header or footer repeated on every page.

SkipCause

Why the stream was skipped.

SkippedTextKind

Which kind of operator stream was skipped.

Traits

Output

Renders laid-out pages into one document.

Functions

document_layout

Every page’s spans as structure, ranking heading sizes against the whole document, so one oversized page cannot redefine what body text is.

document_layout_with_rulings

document_layout with each page’s rulings, so drawn grids become tables document-wide. With no rulings the two are the same function.

extract_markdown

Extracts the whole document as Markdown: ATX headings, paragraphs, and emphasis over the same positional layout extract_text renders flat.

extract_markdown_reporting

extract_markdown with one ExtractReport per page, in page order.

extract_page_markdown

One page as Markdown, ranking heading sizes against that page alone. extract_markdown is the better answer whenever the document is at hand — a page whose text is all one size has no heading to find.

extract_page_markdown_with

extract_page_markdown against any object source. Signed like extract_text_with, for the same reasons — oc gating included.

extract_text

Extracts the page’s text with positional layout applied: spans grouped into lines, lines ordered top to bottom and joined with \n, spaces inserted at horizontal gaps.

extract_text_reporting

extract_text with the report of what could not be read: an ExtractReport whose entries name each skipped stream and why — unsupported filters (the passthrough image codecs included), undecodable bytes, unparseable content, missing resources, exhausted form limits. An empty text with an empty report really is an empty page.

extract_text_reporting_cached

extract_text_reporting with fonts cached across pages: a caller walking a whole document — pdfboss_core::map_pages included — passes one FontCache to every page and each font loads once for the document. The text is identical to the uncached call’s, page for page.

extract_text_reporting_with

extract_text_reporting against any object source. Signed like extract_text_with, for the same reasons — oc gating included.

extract_text_with

extract_text against any object source, awaiting whatever I/O the source needs to read the page — the same span extraction and layout. oc is the document’s optional-content visibility (the async document’s oc_state()); None extracts every layer.

layout

Groups spans into lines (baselines within 0.5 · size), orders lines top to bottom and spans left to right, inserts a space at horizontal gaps wider than WORD_GAP times the size, and joins lines with \n. A page with a clear two-column gutter reads column-major: full-width separators split it into bands, and within each band the left column flows before the right.

page_layout

The page’s spans as structure, ranking heading sizes against this page alone. Prefer document_layout whenever the whole document is at hand: a page of nothing but large type has no body size of its own.

page_layout_with_rulings

page_layout with the page’s rulings: a lattice of drawn borders is read as a table ahead of lane occupancy. With no rulings the two are the same function.