Method load

src/api/pdf.ts:384–472

(bytes: Uint8Array, options?: LoadOptions)

   */
  // oxlint-disable-next-line typescript/require-await
  static async load(bytes: Uint8Array, options?: LoadOptions): Promise<PDF> {
    const scanner = new Scanner(bytes);
    const parser = new DocumentParser(scanner, options);

    const parsed = parser.parse();

    // Create registry from xref
    const registry = new ObjectRegistry(parsed.xref);

    // Detect linearization by checking first object
    let isLinearized = false;

    try {
      // Find the first object (lowest object number)
      const firstObjNum = min(parsed.xref.keys(), 0);

      if (firstObjNum > 0) {
        const firstObj = parsed.getObject(PdfRef.of(firstObjNum, 0));

        if (firstObj instanceof PdfDict && isLinearizationDict(firstObj)) {
          isLinearized = true;
        }
      }
    } catch {
      // Ignore errors during linearization detection
    }

    // Find original xref offset and detect format
    const xrefParser = new XRefParser(scanner);
    let originalXRefOffset: number;
    let usesXRefStreams = false;

    try {
      originalXRefOffset = xrefParser.findStartXRef();
      // Detect if the original uses XRef streams
      const format = xrefParser.detectXRefFormat(originalXRefOffset);
      usesXRefStreams = format === true;
    } catch {
      originalXRefOffset = 0;
    }
    // Set up resolver on registry before loading anything
    registry.setResolver(ref => parsed.getObject(ref));

    // Load catalog through registry so it's tracked for changes
    const rootRef = parsed.trailer.getRef("Root");

    if (!rootRef) {
      throw new Error("Document has no catalog (missing /Root in trailer)");
    }

    const catalogDict = registry.resolve(rootRef);

    if (!catalogDict || !(catalogDict instanceof PdfDict)) {
      throw new Error("Document has no catalog");
    }

    const pdfCatalog = new PDFCatalog(catalogDict, registry);
    const pagesRef = catalogDict.getRef("Pages");