Class XlsxParser

apps/sim/lib/file-parsers/xlsx-parser.ts:20–208

}

export class XlsxParser implements FileParser {
  /**
   * Read the file into a buffer and delegate to {@link parseBuffer} so the
   * decompression-bomb guard runs before SheetJS inflates the workbook.
   */
  async parseFile(filePath: string): Promise<FileParseResult> {
    try {
      if (!filePath) {
        throw new Error('No file path provided')
      }

      if (!existsSync(filePath)) {
        throw new Error(`File not found: ${filePath}`)
      }

      logger.info(`Parsing XLSX file: ${filePath}`)

      const buffer = await readFile(filePath)
      return this.parseBuffer(buffer)
    } catch (error) {
      logger.error('XLSX file parsing error:', error)
      throw new Error(`Failed to parse XLSX file: ${(error as Error).message}`)
    }
  }

  async parseBuffer(buffer: Buffer): Promise<FileParseResult> {
    try {
      const bufferSize = buffer.length
      logger.info(
        `Parsing XLSX buffer, size: ${bufferSize} bytes (${(bufferSize / 1024 / 1024).toFixed(2)} MB)`
      )

      if (!buffer || buffer.length === 0) {
        throw new Error('Empty buffer provided')
      }

      assertOoxmlArchiveWithinLimits(buffer)

      const workbook = XLSX.read(buffer, {
        type: 'buffer',
        dense: true, // Use dense mode for better memory efficiency
        sheetStubs: false, // Don't create stub cells
      })

      return this.processWorkbook(workbook)
    } catch (error) {
      logger.error('XLSX buffer parsing error:', error)
      throw new Error(`Failed to parse XLSX buffer: ${(error as Error).message}`)
    }
  }

  private processWorkbook(workbook: XLSX.WorkBook): FileParseResult {
    const sheetNames = workbook.SheetNames
    let content = ''
    let totalRows = 0
    let truncated = false
    let contentSize = 0
    const sampledData: any[] = []