extract_one_page/
extract_one_page.rs

1//! Extract one page of a PDF.
2
3use hayro_syntax::Pdf;
4use hayro_write::{
5    ChunkSettings, ExtractionQuery, extract,
6    pdf_writer::{self, Ref},
7};
8use std::path::PathBuf;
9
10fn main() {
11    // First load the data that constitutes the PDF file.
12    let data = std::fs::read(PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("../in.pdf")).unwrap();
13    let page_index = 0;
14
15    // Then create a new PDF file from it.
16    //
17    // Here we are just unwrapping in case reading the file failed, but you
18    // might instead want to apply proper error handling.
19    let pdf = Pdf::new(data).unwrap();
20
21    let query = ExtractionQuery::new_page(page_index);
22
23    let mut counter = 0;
24    let mut new_ref = || {
25        counter += 1;
26        Ref::new(counter)
27    };
28    let extraction_result = extract(
29        &pdf,
30        Box::new(&mut new_ref),
31        ChunkSettings { pretty: false },
32        &[query],
33    )
34    .unwrap();
35
36    let chunk = extraction_result.chunk;
37    let page_ref = extraction_result.root_refs[0].unwrap();
38    let parent = extraction_result.page_tree_parent_ref;
39
40    let catalog_id = new_ref();
41    let mut out_pdf = pdf_writer::Pdf::with_settings(ChunkSettings { pretty: false });
42    out_pdf.extend(&chunk);
43    out_pdf.catalog(catalog_id).pages(parent);
44    out_pdf.pages(parent).kids([page_ref]).count(1);
45
46    std::fs::write(
47        PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("../out.pdf"),
48        out_pdf.finish(),
49    )
50    .unwrap();
51}
extract_one_page/extract_one_page.rs

extract_one_page/
extract_one_page.rs
