PDF output

PdfRenderer is a hand-written PDF writer with subset font embedding, searchable text and a ToUnicode map built from the shaping cluster map. Its output was compared operator for operator with LibreOffice's own PDF export on a set of documents in four formats.

What the file carries

Render with the overload that takes the document, so the file is written with the document's own facts:

using XRay.Office;
using XRay.Office.Core.Documents;
using XRay.Office.Rendering.Pdf;

using IDocument doc = OfficeDocument.Open("report.docx", token);
IPageSequence pages = ((IPaginatedDocument)doc).Layout(null, token);

using FileStream output = File.Create("report.pdf");
new PdfRenderer(new PdfRenderOptions { Language = "en-US" }).Render(doc, pages, output, token);

The PDF has:

Document information and XMP: the title, author, subject, keywords and the application that made the source, from IDocument.Metadata, in the information dictionary and in an XMP packet.
Language: the catalogue's /Lang, from PdfRenderOptions.Language or the document's own.
Links: a /Link annotation with a /URI action for every hyperlink, and annotations with a destination for every link to a bookmark, cross-reference or table-of-contents entry, with a name tree of the destinations.
An outline: the bookmarks panel, built from the headings of a word-processing document, the sheets of a workbook and the slides of a deck. A slide is called by its name when the deck gives one, else by the first line of its text, else "Slide N".
Page labels, where the printed page numbers differ from the pages' positions, as after a title page or a restarted section.
A file identifier (/ID).

Links, outline and labels are declared by the layout to the renderer, from DOCX, ODT, RTF and DOC documents, from sheet names, and from slide titles.

Options

`PdfRenderOptions`	Default	Effect
`WriteDocumentInfo`	`true`	The information dictionary.
`WriteXmp`	`true`	The XMP packet.
`WriteLinks`	`true`	Link annotations and destinations.
`WriteOutline`	`true`	The bookmarks panel.
`WritePageLabels`	`true`	`/PageLabels`.
`Language`	`null`	The `/Lang` tag; taken from the document when `null`.
`Metadata`	`null`	The document metadata to write; set automatically by the overloads that take an `IDocument`.
`EmbedFonts`	`true`	Embed subset fonts.
`PassThroughJpeg`	`true`	Copy JPEG pictures without re-encoding.
`CreationDate`	`null`	A fixed timestamp for reproducible output.
`Producer`	`X-Ray.Office`	The producer string.
`MaxImagePixels`, `OversizedImages`	`0`, `Skip`	Picture limits; see Cancellation and limits.
`FontProvider`	file fonts	Where the bytes of a face come from.

The overload Render(pages, stream) without a document writes the producer and date only, plus whatever you put in Metadata.

Other outputs

RasterRenderer writes PNG (lossless and deterministic) and JPEG at a chosen DPI over SkiaSharp. SheetHtmlWriter writes a workbook as HTML. MarkdownWriter and XhtmlWriter write the extracted content tree as GitHub-flavoured Markdown and XHTML.