Library doesn't preserve PDF outlines (Table of Contents)

[yusifaliyevpro]

When processing a PDF with @libpdf/core, the resulting output file does not retain the original document outlines. Outlines - internal navigation links.

[Mythie]

This seems weird, while the library doesn't support modifying outlines at the moment it should preserve them provided they're reachable.

Do you have the PDF on hand that is getting them omitted.

[yusifaliyevpro]

@Mythie, I have a PDF for testing. The issue only occurs with this code. Other implementations like watermarking, protecting, and unlocking PDF, preserve the table of contents.

8e1ecf55-f45e-4f6d-822e-9504deae7f60.pdf

/**
 * Props for splitting by size on Windows
 */
type SplitBySizePropsWindows = {
  pdf: PDF;
  maxSizeBytes: number;
  originalFileName: string;
  dirPath: string;
  suffix: string;
};
/**
 * For Windows, MacOS uses native Swift implementation.
 * Splits a PDF document by file size, creating chunks that don't exceed the specified size limit.
 */
export async function splitBySizeWindows(props: SplitBySizePropsWindows): Promise<void> {
  const { pdf, maxSizeBytes, originalFileName, dirPath, suffix } = props;
  const totalPages = pdf.getPageCount();
  let currentPart = 1;
  let startPage = 0;

  while (startPage < totalPages) {
    let endPage = startPage;
    let currentChunkBytes: Uint8Array | null = null;

    // Greedily add pages until we exceed the size limit
    while (endPage < totalPages) {
      const indices: number[] = [];
      for (let i = startPage; i <= endPage; i++) {
        indices.push(i);
      }

      const chunkPdf = await pdf.extractPages(indices);
      const testBytes = await chunkPdf.save();

      if (testBytes.length > maxSizeBytes) {
        // If this is the first page and it already exceeds the limit,
        // we must include it anyway (single page exceeds max size)
        if (endPage === startPage) {
          currentChunkBytes = testBytes;
          endPage++;
        }
        break;
      }

      currentChunkBytes = testBytes;
      endPage++;
    }

    // Save the current chunk
    if (currentChunkBytes) {
      const chunkFilePath = path.join(dirPath, `${originalFileName} [${suffix} ${currentPart}].pdf`);
      await fs.writeFile(chunkFilePath, currentChunkBytes);
      currentPart++;
    }

    startPage = endPage;
  }
}

[Mythie]

Oh for splitting that would make sense since the outline would point to stuff that no longer exists.

E.g. a table of contents would say page 85 when ur viewing pages 20-25

[yusifaliyevpro]

But I still can't see the existing page titles. At least, I should be able to view exisiting ones. iLovePDF correctly splits by size on their website.

[Mythie]

Right yeah, truthfully the library has nothing for outlines at the moment. Shouldn't be hard to add it though, i'll take a look over the weekend if I've got some spare time 😄

[yusifaliyevpro]

Thanks. No rush. Let me know when it is ready and published on npm

[symblify]

Support for outlines would be a great addition and much appreciated

[xilopaint]

Right yeah, truthfully the library has nothing for outlines at the moment. Shouldn't be hard to add it though, i'll take a look over the weekend if I've got some spare time 😄

Hey @Mythie, is this expected to be implemented in the near future?