Class PptxParser

apps/sim/lib/file-parsers/pptx-parser.ts:10–109

const logger = createLogger('PptxParser')

export class PptxParser implements FileParser {
  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 PowerPoint file: ${filePath}`)

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

  async parseBuffer(buffer: Buffer): Promise<FileParseResult> {
    try {
      logger.info('Parsing PowerPoint buffer, size:', buffer.length)

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

      assertOoxmlArchiveWithinLimits(buffer)

      let parseOfficeAsync
      try {
        const officeParser = await import('officeparser')
        parseOfficeAsync = officeParser.parseOfficeAsync
      } catch (importError) {
        logger.warn('officeparser not available, using fallback extraction')
        return this.fallbackExtraction(buffer)
      }

      try {
        const result = await parseOfficeAsync(buffer)

        if (!result || typeof result !== 'string') {
          throw new Error('officeparser returned invalid result')
        }

        const content = sanitizeTextForUTF8(result.trim())

        logger.info('PowerPoint parsing completed successfully with officeparser')

        return {
          content: content,
          metadata: {
            characterCount: content.length,
            extractionMethod: 'officeparser',
          },
        }