Export Markdown.gs

/*
Credit:
  The vast majority of this code comes from:
  https://github.com/mangini/gdocs2md/issues
  
  Ben Jefferson added the following to create this fork
  - Replaced emailing of markdown files with creating files in Google Drive in a "markdown" directory alongside the file being converted.
  - Added support for linking to (and automatic conversion) other Google docs in the same directory or a subdirectory
License:
  The original file and Ben Jefferson's subsequent edits are all released under the Apache2 License see https://github.com/mangini/gdocs2md for more details
  
Usage: 
  Adding this script to your doc: 
    - Tools > Script Manager > New
    - Select "Blank Project", then paste this code in and save as "Export Markdown.gs"
    - Run > Test as add-on
    - Choose the document you want to test this on - this will open the document in a new tab
  Running the script:
    - Make sure there is a "markdown" directory in the same directory as the document you are editting - this is where the generated files will go
    - Add-ons > Convert To Markdown > update Markdown
    - On successfull completion a pop-up will appear saying "All Done"
*/

var outputDirSuffix='_markdown';

var parentFolder;
var outputBaseDir;
var processedFiles;
var usedFilenames;

function onInstall(e) {
  onOpen(e);
}

function onOpen(e) {
  DocumentApp.getUi().createAddonMenu()
      .addItem('Update Markdown', 'processActiveDocument')
      .addToUi();
}

function processActiveDocument() {
  processedFiles = {};
  usedFilenames = {};
  var theDocument = DocumentApp.getActiveDocument();
  parentFolder = DriveApp.getFileById(theDocument.getId()).getParents();
  // A file can live in multiple directories - this code just uses the first
  parentFolder = parentFolder.next();
  
  var outputBaseDirName = theDocument.getName() + outputDirSuffix;
  outputBaseDir = parentFolder.getFoldersByName(outputBaseDirName);

  // Create the output dir if it doesn't already exist
  if (outputBaseDir.hasNext()) outputBaseDir = outputBaseDir.next();
  else outputBaseDir = parentFolder.createFolder(outputBaseDirName);

  ConvertToMarkdown( theDocument );
  
  // Remove any old unused files from the output directory
  var outputDirFiles = outputBaseDir.getFiles();
  
  while (outputDirFiles.hasNext()) {
    var file = outputDirFiles.next();
    if (!usedFilenames[file.getName()]) outputBaseDir.removeFile(file);
  }
  
  DocumentApp.getUi().alert('All done!');
}

function getFilename( filename, extension ){
  var count=0;
  var testFilename;
  do {
    testFilename = filename;
    if (count) testFilename += '_'+count;
    if (extension) testFilename += extension;
    count++;
  } while (usedFilenames[testFilename]);
  usedFilenames[testFilename] = true;
  var needsDeleting = outputBaseDir.getFilesByName(testFilename);
  if (needsDeleting.hasNext()) {
    outputBaseDir.removeFile(needsDeleting.next());
  }
  return testFilename;
}

function ConvertToMarkdown( theDocument ) {
  if (processedFiles[theDocument.getId()]) return processedFiles[theDocument.getId()];
  
  var processedFilename = getFilename(theDocument.getName(),".md");
  processedFiles[theDocument.getId()]=processedFilename;
  var imageBaseFilename = processedFilename.slice(0,-3);
  
  
  var body = theDocument.getBody();
  var numChildren = body.getNumChildren();
  var text = "";
  var inSrc = false;
  var inClass = false;
  var globalListCounters = {};
  // edbacher: added a variable for indent in src <pre> block. Let style sheet do margin.
  var srcIndent = "";
    
  // Walk through all the child elements of the doc.
  for (var i = 0; i < numChildren; i++) {
    var child = body.getChild(i);
    var result = processParagraph(i, child, inSrc, imageBaseFilename, globalListCounters);
    if (result!==null) {
      if (result.sourcePretty==="start" && !inSrc) {
        inSrc=true;
        text+="<pre class=\"prettyprint\">\n";
      } else if (result.sourcePretty==="end" && inSrc) {
        inSrc=false;
        text+="</pre>\n\n";
      } else if (result.source==="start" && !inSrc) {
        inSrc=true;
        text+="<pre>\n";
      } else if (result.source==="end" && inSrc) {
        inSrc=false;
        text+="</pre>\n\n";
      } else if (result.inClass==="start" && !inClass) {
        inClass=true;
        text+="<div class=\""+result.className+"\">\n";
      } else if (result.inClass==="end" && inClass) {
        inClass=false;
        text+="</div>\n\n";
      } else if (inClass) {
        text+=result.text+"\n\n";
      } else if (inSrc) {
        text+=(srcIndent+escapeHTML(result.text)+"\n");
      } else if (result.text && result.text.length>0) {
        text+=result.text+"\n\n";
      }
      
      if (result.images && result.images.length>0) {
        for (var j=0; j<result.images.length; j++) {
          createOutputFile(outputBaseDir,getFilename(result.images[j].name), result.images[j].bytes, result.images[j].type, true );
        }
      }
    } else if (inSrc) { // support empty lines inside source code
      text+='\n';
    }
      
  }
  
  createOutputFile(outputBaseDir,processedFilename, text, "text/plain" );
  
  return processedFilename;
}

function createOutputFile( outputBaseDir, filename, content, type, isBlob ) {
  var existingFiles = outputBaseDir.getFilesByName( filename );
  while (existingFiles.hasNext()) {
    outputBaseDir.removeFile(existingFiles.next());
  }
  outputBaseDir.createFile(Utilities.newBlob(content, type, filename));
}

function escapeHTML(text) {
  return text.replace(/</g, '&lt;').replace(/>/g, '&gt;');
}

// Process each child element (not just paragraphs).
function processParagraph(index, element, inSrc, imageBaseFilename, listCounters) {
  // First, check for things that require no processing.
  if (element.getNumChildren()==0) {
    return null;
  }  
  // Punt on TOC.
  if (element.getType() === DocumentApp.ElementType.TABLE_OF_CONTENTS) {
    return {"text": "[[TOC]]"};
  }
  
  // Set up for real results.
  var result = {};
  var pOut = "";
  var textElements = [];
  var imageCounter = 0;
  
  // Handle Table elements. Pretty simple-minded now, but works for simple tables.
  // Note that Markdown does not process within block-level HTML, so it probably 
  // doesn't make sense to add markup within tables.
  if (element.getType() === DocumentApp.ElementType.TABLE) {
    textElements.push("<table>\n");
    var nCols = element.getChild(0).getNumCells();
    for (var i = 0; i < element.getNumChildren(); i++) {
      textElements.push("  <tr>\n");
      // process this row
      for (var j = 0; j < nCols; j++) {
        textElements.push("    <td>" + element.getChild(i).getChild(j).getText() + "</td>\n");
      }
      textElements.push("  </tr>\n");
    }
    textElements.push("</table>\n");
  }
  
  // Process various types (ElementType).
  for (var i = 0; i < element.getNumChildren(); i++) {
    var t=element.getChild(i).getType();
    
    if (t === DocumentApp.ElementType.TABLE_ROW) {
      // do nothing: already handled TABLE_ROW
    } else if (t === DocumentApp.ElementType.TEXT) {
      var txt=element.getChild(i);
      pOut += txt.getText();
      textElements.push(txt);
    } else if (t === DocumentApp.ElementType.INLINE_IMAGE) {
      result.images = result.images || [];
      var contentType = element.getChild(i).getBlob().getContentType();
      var extension = "";
      if (/\/png$/.test(contentType)) {
        extension = ".png";
      } else if (/\/gif$/.test(contentType)) {
        extension = ".gif";
      } else if (/\/jpe?g$/.test(contentType)) {
        extension = ".jpg";
      } else {
        throw "Unsupported image type: "+contentType;
      }
      var name = imageBaseFilename + '_' + imageCounter + extension;
      imageCounter++;
      textElements.push('![image alt text]('+name+')');
      result.images.push( {
        "bytes": element.getChild(i).getBlob().getBytes(), 
        "type": contentType, 
        "name": name});
    } else if (t === DocumentApp.ElementType.PAGE_BREAK) {
      // ignore
    } else if (t === DocumentApp.ElementType.HORIZONTAL_RULE) {
      textElements.push('* * *\n');
    } else if (t === DocumentApp.ElementType.FOOTNOTE) {
      textElements.push(' (NOTE: '+element.getChild(i).getFootnoteContents().getText()+')');
    } else {
      throw "Paragraph "+index+" of type "+element.getType()+" has an unsupported child: "
      +t+" "+(element.getChild(i)["getText"] ? element.getChild(i).getText():'')+" index="+index;
    }
  }

  if (textElements.length==0) {
    // Isn't result empty now?
    return result;
  }
  
  // evb: Add source pretty too. (And abbreviations: src and srcp.)
  // process source code block:
  if (/^\s*---\s+srcp\s*$/.test(pOut) || /^\s*---\s+source pretty\s*$/.test(pOut)) {
    result.sourcePretty = "start";
  } else if (/^\s*---\s+src\s*$/.test(pOut) || /^\s*---\s+source code\s*$/.test(pOut)) {
    result.source = "start";
  } else if (/^\s*---\s+class\s+([^ ]+)\s*$/.test(pOut)) {
    result.inClass = "start";
    result.className = RegExp.$1;
  } else if (/^\s*---\s*$/.test(pOut)) {
    result.source = "end";
    result.sourcePretty = "end";
    result.inClass = "end";
  } else if (/^\s*---\s+jsperf\s*([^ ]+)\s*$/.test(pOut)) {
    result.text = '<iframe style="width: 100%; height: 340px; overflow: hidden; border: 0;" '+
                  'src="http://www.html5rocks.com/static/jsperfview/embed.html?id='+RegExp.$1+
                  '"></iframe>';
  } else {

    var prefix = findPrefix(inSrc, element, listCounters);
  
    var pOut = "";
    for (var i=0; i<textElements.length; i++) {
      pOut += processTextElement(inSrc, textElements[i]);
    }

    // replace Unicode quotation marks
    pOut = pOut.replace('\u201d', '"').replace('\u201c', '"');
 
    result.text = prefix+pOut;
  }
  
  return result;
}

// Add correct prefix to list items.
function findPrefix(inSrc, element, listCounters) {
  var prefix="";
  if (!inSrc) {
    if (element.getType()===DocumentApp.ElementType.PARAGRAPH) {
      var paragraphObj = element;
      switch (paragraphObj.getHeading()) {
        // Add a # for each heading level. No break, so we accumulate the right number.
        case DocumentApp.ParagraphHeading.HEADING6: prefix+="#";
        case DocumentApp.ParagraphHeading.HEADING5: prefix+="#";
        case DocumentApp.ParagraphHeading.HEADING4: prefix+="#";
        case DocumentApp.ParagraphHeading.HEADING3: prefix+="#";
        case DocumentApp.ParagraphHeading.HEADING2: prefix+="#";
        case DocumentApp.ParagraphHeading.HEADING1: prefix+="# ";
        default:
      }
    } else if (element.getType()===DocumentApp.ElementType.LIST_ITEM) {
      var listItem = element;
      var nesting = listItem.getNestingLevel()
      for (var i=0; i<nesting; i++) {
        prefix += "    ";
      }
      var gt = listItem.getGlyphType();
      // Bullet list (<ul>):
      if (gt === DocumentApp.GlyphType.BULLET
          || gt === DocumentApp.GlyphType.HOLLOW_BULLET
          || gt === DocumentApp.GlyphType.SQUARE_BULLET) {
        prefix += "* ";
      } else {
        // Ordered list (<ol>):
        var key = listItem.getListId() + '.' + listItem.getNestingLevel();
        var counter = listCounters[key] || 0;
        counter++;
        listCounters[key] = counter;
        prefix += counter+". ";
      }
    }
  }
  return prefix;
}

function processLinkedFile( url ) {
  if (url.substr(0,24) !== 'https://docs.google.com/') return url;
  var fileId = url.match(/.*[^-\w]([-\w]{25,})[^-\w]?.*/);
  if (!fileId || !fileId[1]) return url;
  fileId = fileId[1];
  
  linkedFile = DriveApp.getFileById(fileId);
  if (!linkedFile) return url;
  
  if (linkedFile.getMimeType()!=MimeType.GOOGLE_DOCS) return url
  
  newUrl = ConvertToMarkdown( DocumentApp.openById(fileId) );  
  if (newUrl.length) return newUrl;
  
  return url;
}

function processTextElement(inSrc, txt) {
  if (typeof(txt) === 'string') {
    return txt;
  }
  
  var pOut = txt.getText();
  if (! txt.getTextAttributeIndices) {
    return pOut;
  }
  
  var attrs=txt.getTextAttributeIndices();
  var lastOff=pOut.length;

  for (var i=attrs.length-1; i>=0; i--) {
    var off=attrs[i];
    var url=txt.getLinkUrl(off);
    var font=txt.getFontFamily(off);
    if (url) {  // start of link
      if (i>=1 && attrs[i-1]==off-1 && txt.getLinkUrl(attrs[i-1])===url) {
        // detect links that are in multiple pieces because of errors on formatting:
        i-=1;
        off=attrs[i];
        url=txt.getLinkUrl(off);
      }
      
      url = processLinkedFile( url );
      pOut=pOut.substring(0, off)+'['+pOut.substring(off, lastOff)+']('+url+')'+pOut.substring(lastOff);
    } else if (font) {
      if (!inSrc && font===font.COURIER_NEW) {
        while (i>=1 && txt.getFontFamily(attrs[i-1]) && txt.getFontFamily(attrs[i-1])===font.COURIER_NEW) {
          // detect fonts that are in multiple pieces because of errors on formatting:
          i-=1;
          off=attrs[i];
        }
        pOut=pOut.substring(0, off)+'`'+pOut.substring(off, lastOff)+'`'+pOut.substring(lastOff);
      }
    }
    if (txt.isBold(off)) {
      var d1 = d2 = "**";
      if (txt.isItalic(off)) {
        // edbacher: changed this to handle bold italic properly.
        d1 = "**_"; d2 = "_**";
      }
      pOut=pOut.substring(0, off)+d1+pOut.substring(off, lastOff)+d2+pOut.substring(lastOff);
    } else if (txt.isItalic(off)) {
      pOut=pOut.substring(0, off)+'*'+pOut.substring(off, lastOff)+'*'+pOut.substring(lastOff);
    }
    lastOff=off;
  }
  return pOut;
}