pub fn pdf_words(bytes: &[u8]) -> Vec<(f32, f32, Vec<TextCell>)>Expand description
Debug/diagnostic entry: per-page (width, height, word cells) for a PDF, via
the Rust parser glyphs run through the docling-parse word grouping. Used to
compare parser word cells against docling-parse’s word_cells oracle (roadmap
item 6).