Expand description
Extract highlighted text and comments from an annotated PDF.
Text markup annotations (Highlight/Underline/StrikeOut/Squiggly) carry /QuadPoints but not the text they cover, so that text is recovered by intersecting each quad with the page’s glyph boxes. Text and FreeText annotations have no quads and yield comment-only records.
Three pdf_oxide behaviours are worked around here:
- /Contents is decoded with from_utf8_lossy, which mangles the UTF-16BE
strings most annotators write. We re-decode from
raw_dict. - The glyph coordinate convention is not reliably documented, so rather
than trusting a doc comment we infer it from the glyphs themselves
(see
GlyphSpace::infer) and show the evidence under –debug-geometry. - /ActualText is re-encoded through the font in character mode, which
corrupts every declared span. See
actual_text.
Re-exports§
pub use markdown::BareStyle;pub use markdown::Descriptor;pub use markdown::Escaping;pub use markdown::Numbering;pub use markdown::Section;pub use markdown::Style;pub use diagnostic::Diagnostic;pub use diagnostic::Diagnostics;pub use diagnostic::Severity;pub use error::Error;
Modules§
- diagnostic
- What had to be assumed, worked around, or given up on.
- error
- The four ways a report fails to be produced.
- markdown
- Markdown rendering.
Structs§
Enums§
Functions§
- extract
- Read every annotation worth reporting out of one PDF.