Class DocumentParser

src/parser/document-parser.ts:143–913

 * ```
 */
export class DocumentParser {
  private readonly scanner: Scanner;
  private readonly options: ParseOptions;
  private readonly warnings: string[] = [];

  constructor(scanner: Scanner, options: ParseOptions = {}) {
    this.scanner = scanner;
    this.options = {
      lenient: options.lenient ?? true,
      credentials: options.credentials,
    };
  }

  /**
   * Parse the PDF document.
   */
  parse(): ParsedDocument {
    try {
      return this.parseNormal();
    } catch (error) {
      // Only attempt recovery for recoverable parsing errors
      if (this.options.lenient && error instanceof RecoverableParseError) {
        this.warnings.push(`Normal parsing failed: ${error.message}`);

        return this.parseWithRecovery();
      }

      // All other errors propagate (credentials, unsupported features, etc.)
      throw error;
    }
  }

  /**
   * Normal parsing path.
   */
  private parseNormal(): ParsedDocument {
    // Phase 1: Parse header
    const version = this.parseHeader();

    // Phase 2: Find startxref
    const xrefParser = new XRefParser(this.scanner);

    const startXRef = xrefParser.findStartXRef();

    // Phase 3: Parse XRef chain (follow /Prev)
    const { xref, trailer } = this.parseXRefChain(xrefParser, startXRef);

    // Phase 4: Build document with lazy object loading
    return this.buildDocument(version, xref, trailer, false);
  }

  /**
   * Recovery parsing using brute-force when normal parsing fails.
   */
  private parseWithRecovery(): ParsedDocument {
    // Try to get version even if header is malformed
    let version = DEFAULT_VERSION;