Annotate Existing Text

Add an Underline or StrikeOut annotation to text already present in an unrotated PDF by extracting its text-showing operations, converting the PDF bottom-left origin to Recipe's top-left coordinates, then editing the page.

This example marks every text operation that reads exactly Draft. Matching uses the decoded text, so non-ASCII words work too. width must describe the known visual bounds in the source template; use a separate region for each different width.

var muhammara = require("@muhammara/native");
var Recipe = muhammara.Recipe;

/**
 * Underlines every text-showing operation on an unrotated page whose decoded
 * text is `text`.
 *
 * @param {string} input Source PDF path.
 * @param {string} output Output PDF path.
 * @param {number} pageNumber One-based page number.
 * @param {string} text Text to mark.
 * @param {number} width Known visual width of that text in the template.
 * @returns {number} The number of marked text operations.
 */
function annotateExistingText(input, output, pageNumber, text, width) {
  var pageIndex = pageNumber - 1; // Reader pages are zero-based.
  var reader = muhammara.createReader(input);
  var matches;

  try {
    matches = reader.extractPageText(pageIndex).filter(function (element) {
      return element.text === text;
    });
  } finally {
    reader.end();
  }

  var pdf = new Recipe(input, output);
  var page = pdf.pageInfo(pageNumber);

  pdf.editPage(pageNumber);
  matches.forEach(function (match) {
    pdf.annot(
      match.textMatrix[4],
      page.height - match.textMatrix[5] - match.fontSize,
      Recipe.AnnotSubtype.UNDERLINE,
      {
        width: width,
        height: match.fontSize,
        color: "#008000",
        text: "Reviewed",
      },
    );
  });
  pdf.endPage().endPDF();
  return matches.length;
}

annotateExistingText("input.pdf", "reviewed.pdf", 1, "Draft", 28);

Use Recipe.AnnotSubtype.STRIKE_OUT instead to create strikeout annotations. The reader's textMatrix uses PDF's bottom-left coordinates, while Recipe uses a top-left origin; the page.height - y - fontSize conversion above applies to unrotated pages only.

extractPageText() reports a text operation's decoded text, raw content, origin, font resource, and font size. It does not calculate glyph bounds, so it cannot accurately derive width for arbitrary PDFs. This pattern is appropriate for a controlled template where the target text and its bounds are known. For arbitrary documents, obtain glyph bounds from another layout or text-analysis tool before creating the annotation. See Find Text Positions In A PDF for the extractor's complete limitations, and Add Review Annotations for annotation options.