如何用Google Apps Script翻译Google Docs文件并保留格式、表格与图片？

问题背景 #

我需要用Google Apps Script实现类似Google Docs自带“翻译文档”的功能，批量翻译指定文件夹内的文档。但使用LanguageApp.translate()提取整个文档纯文本翻译时，会丢失表格边框、字体样式等所有格式，现有代码如下：

function TranslateFunction() {
  //Get the files in your indicated Folder
  var TargetFolderID = '1VUNGtqiNbnHhIFCXmbdSwNZ-vZ5NWVTE'; //Paste the folder ID here to start
  var folder = DriveApp.getFolderById(TargetFolderID);
  var files = folder.getFiles();
  //Get all the files' ID in the folder above
  while (files.hasNext()){
    var file = files.next();
    var fileID = file.getId();

    //Convert each file in the folder from Docx (Word) to Docs (Google)
    var docx = DriveApp.getFileById(fileID);
    var newDoc = Drive.newFile();
    var blob = docx.getBlob();
    var file=Drive.Files.insert(newDoc,blob,{convert:true});
    DocumentApp.openById(file.id).setName(docx.getName().slice(0,-5));  

    //Activate the file
    var doc = DocumentApp.openById(file.id);

    //Create a new Docs file to put the translation in + Name
    var newDoc = DocumentApp.create("|EN| " + docx.getName().slice(0,-5)); 
  
    //Get the text in the file
    var bodyText = doc.getBody().getText();

    //Translate the text and append it into the new Docs
    Translatedtext = LanguageApp.translate(bodyText,'vi','en');
    newDoc.getBody().appendParagraph(Translatedtext);
  }
}

解决思路与代码 #

要保留格式，必须逐元素处理文档内容，翻译每个元素的文本同时复制原格式。以下是修改后的代码，支持段落、表格、标题等常见元素的格式保留：

function batchTranslateDocsWithFormat() {
  const targetFolderId = '1VUNGtqiNbnHhIFCXmbdSwNZ-vZ5NWVTE'; // 目标文件夹ID
  const sourceLang = 'vi'; // 源语言
  const targetLang = 'en'; // 目标语言
  const folder = DriveApp.getFolderById(targetFolderId);
  const files = folder.getFiles();

  while (files.hasNext()) {
    const file = files.next();
    const fileName = file.getName();
    
    // 仅处理docx格式文件
    if (!fileName.endsWith('.docx')) continue;

    // 将docx转换为Google Docs格式
    const blob = file.getBlob();
    const convertedFile = Drive.Files.insert({}, blob, {convert: true});
    const convertedDoc = DocumentApp.openById(convertedFile.id);
    convertedDoc.setName(fileName.slice(0, -5));

    // 创建翻译后的新文档
    const translatedDoc = DocumentApp.create(`|EN| ${fileName.slice(0, -5)}`);
    const sourceBody = convertedDoc.getBody();
    const targetBody = translatedDoc.getBody();

    // 遍历源文档所有元素，逐个翻译并保留格式
    const numElements = sourceBody.getNumChildren();
    for (let i = 0; i < numElements; i++) {
      const element = sourceBody.getChild(i);
      translateElement(element, targetBody, sourceLang, targetLang);
    }

    // 保存并关闭文档
    translatedDoc.saveAndClose();
    convertedDoc.saveAndClose();
  }
}

// 递归处理不同类型的文档元素
function translateElement(element, targetBody, sourceLang, targetLang) {
  const elementType = element.getType();

  switch(elementType) {
    // 处理普通段落
    case DocumentApp.ElementType.PARAGRAPH:
      const sourcePara = element.asParagraph();
      const targetPara = targetBody.appendParagraph('');
      // 复制段落整体格式
      targetPara.setAttributes(sourcePara.getAttributes());
      // 翻译段落内的每个文本片段，保留局部样式
      const numChildren = sourcePara.getNumChildren();
      for (let j = 0; j < numChildren; j++) {
        const child = sourcePara.getChild(j);
        if (child.getType() === DocumentApp.ElementType.TEXT) {
          const text = child.asText().getText();
          const translatedText = LanguageApp.translate(text, sourceLang, targetLang);
          const targetText = targetPara.appendText(translatedText);
          // 复制文本的字体、颜色等样式
          targetText.setAttributes(child.asText().getAttributes());
        } else {
          // 段落内的非文本元素（如图片）直接复制
          targetPara.appendElement(child.copy());
        }
      }
      break;

    // 处理表格
    case DocumentApp.ElementType.TABLE:
      const sourceTable = element.asTable();
      const targetTable = targetBody.appendTable();
      // 复制表格整体样式（边框、对齐等）
      targetTable.setAttributes(sourceTable.getAttributes());
      const numRows = sourceTable.getNumRows();
      for (let rowIdx = 0; rowIdx < numRows; rowIdx++) {
        const sourceRow = sourceTable.getRow(rowIdx);
        const targetRow = targetTable.appendTableRow();
        const numCells = sourceRow.getNumCells();
        for (let cellIdx = 0; cellIdx < numCells; cellIdx++) {
          const sourceCell = sourceRow.getCell(cellIdx);
          const targetCell = targetRow.appendTableCell();
          // 复制单元格格式
          targetCell.setAttributes(sourceCell.getAttributes());
          // 递归处理单元格内的元素
          const cellChildren = sourceCell.getNumChildren();
          for (let k = 0; k < cellChildren; k++) {
            translateElement(sourceCell.getChild(k), targetCell, sourceLang, targetLang);
          }
        }
      }
      break;

    // 处理标题和列表项
    case DocumentApp.ElementType.HEADING_1:
    case DocumentApp.ElementType.HEADING_2:
    case DocumentApp.ElementType.HEADING_3:
    case DocumentApp.ElementType.LIST_ITEM:
      const sourceHeading = element.asParagraph();
      const targetHeading = targetBody.appendParagraph('');
      targetHeading.setAttributes(sourceHeading.getAttributes());
      const headingText = sourceHeading.getText();
      const translatedHeading = LanguageApp.translate(headingText, sourceLang, targetLang);
      targetHeading.setText(translatedHeading);
      break;

    // 默认处理：直接复制非文本元素（如图片、分隔线）
    default:
      targetBody.appendElement(element.copy());
      break;
  }
}

关键说明 #

• 逐元素处理：不再提取整个文档的纯文本，而是遍历每个元素，翻译文本内容的同时复制原格式
• 递归处理嵌套元素：表格单元格内的段落、文本会被递归处理，确保格式完整保留
• 格式复制：通过getAttributes()和setAttributes()复制元素的样式（字体、颜色、对齐方式、表格边框等）
• 文件过滤：仅处理.docx文件，避免不必要的格式转换

内容的提问来源于stack exchange，提问作者Việt Lương