Method fallbackExtraction

apps/sim/lib/file-parsers/doc-parser.ts:82–129

(buffer: Buffer)

  }

  private fallbackExtraction(buffer: Buffer): FileParseResult {
    const isBinaryDoc = buffer.length >= 2 && buffer[0] === 0xd0 && buffer[1] === 0xcf

    if (!isBinaryDoc) {
      const textContent = buffer.toString('utf8').trim()

      if (textContent.length > 0) {
        const printableChars = textContent.match(/[\x20-\x7E\n\r\t]/g)?.length || 0
        const isProbablyText = printableChars / textContent.length > 0.9

        if (isProbablyText) {
          return {
            content: sanitizeTextForUTF8(textContent),
            metadata: {
              extractionMethod: 'plaintext-fallback',
              characterCount: textContent.length,
              warning: 'File is not a valid DOC format, extracted as plain text',
            },
          }
        }
      }
    }

    const text = buffer.toString('utf8', 0, Math.min(buffer.length, 100000))

    const readableText = text
      .match(/[\x20-\x7E\s]{4,}/g)
      ?.filter(
        (chunk) =>
          chunk.trim().length > 10 && /[a-zA-Z]/.test(chunk) && !/^[\x00-\x1F]*$/.test(chunk)
      )
      .join(' ')
      .replace(/\s+/g, ' ')
      .trim()

    const content = readableText
      ? sanitizeTextForUTF8(readableText)
      : 'Unable to extract text from DOC file. Please convert to DOCX format for better results.'

    return {
      content,
      metadata: {
        extractionMethod: 'fallback',
        characterCount: content.length,
        warning: 'Basic text extraction used. For better results, convert to DOCX format.',
      },
    }
  }
}