【js文件系列三】JS如何提取Pdf 中的图片和文字




从 PDF 中提取文字 -核心代码



 * Retrieves the text of a specif page within a PDF Document obtained through pdf.js
 * @param {Integer} pageNum Specifies the number of the page
 * @param {PDFDocument} PDFDocumentInstance The PDF document obtained
function getPageText(pageNum, PDFDocumentInstance) {
  // Return a Promise that is solved once the text of the page is retrieven
  return new Promise(function (resolve, reject) {
    PDFDocumentInstance.getPage(pageNum).then(function (pdfPage) {
      // The main trick to obtain the text of the PDF page, use the getTextContent method
      pdfPage.getTextContent().then(function (textContent) {
        var textItems = textContent.items;
        var finalString = '';

        // Concatenate the string of the item to the final string
        for (var i = 0; i < textItems.length; i++) {
          var item = textItems[i];

          finalString += item.str + ' ';

        // Solve promise with the text retrieven from the page

从 PDF 中提取图片


// first here I open the document
pdf.getDocument('haorooms.pdf').promise.then(async function (pdfObj) {
  // because I am testing, I just wanted to get page 7
  const page = await pdfObj.getPage(7);

  // now I need to get the image information and for that I get the operator list
  const operators = await page.getOperatorList();

  // this is for the paintImageXObject one, there are other ones, like the paintJpegImage which I assume should work the same way, this gives me the whole list of indexes of where an img was inserted
  const rawImgOperator = operators.fnArray
    .map((f, index) => (f === pdf.OPS.paintImageXObject ? index : null))
    .filter((n) => n !== null);

  // now you need the filename, in this example I just picked the first one from my array, your array may be empty, but I knew for sure in page 7 there was an image... in your actual code you would use loops, such info is in the argsArray, the first arg is the filename, second arg is the width and height, but the filename will suffice here
  const filename = operators.argsArray[rawImgOperator[0]][0];

  // now we get the object itself from page.objs using the filename
  page.objs.get(filename, async (arg) => {
    // and here is where we need the canvas, the object contains information such as width and height
    const canvas = ccc.createCanvas(arg.width, arg.height);
    const ctx = canvas.getContext('2d');

    // now you need a new clamped array because the original one, may not contain rgba data, and when you insert you want to do so in rgba form, I think that a simple check of the size of the clamped array should work, if it's 3 times the size aka width*height*3 then it's rgb and shall be converted, if it's 4 times, then it's rgba and can be used as it is; in my case it had to be converted, and I think it will be the most common case
    const data = new Uint8ClampedArray(arg.width * arg.height * 4);
    let k = 0;
    let i = 0;
    while (i < arg.data.length) {
      data[k] = arg.data[i]; // r
      data[k + 1] = arg.data[i + 1]; // g
      data[k + 2] = arg.data[i + 2]; // b
      data[k + 3] = 255; // a

      i += 3;
      k += 4;

    // now here I create the image data context
    const imgData = ctx.createImageData(arg.width, arg.height);
    ctx.putImageData(imgData, 0, 0);

    // get myself a buffer
    const buff = canvas.toBuffer();

    // and I wrote the file, worked like charm, but this buffer encodes for a png image, which can be rather large, with an image conversion utility like sharp.js you may get better results by compressing the thing.
    fs.writeFile('test', buff);


本文主要介绍了js获取pdf中文本和图片的方法,其实pdf转word也是大致这个思路,主要获取文本和图片,放到word文档中。 本文主要是利用了pdfjs库,参考了issue https://github.com/mozilla/pdf.js/issues/13541