Word processor for selection highlighter meant to be used with GoogleChrome's speechSynthesis engine.

word-processor.js

(function() {
    'use strict';
    /**
     * List of all punctuation marks that are allowed in the sanitized text
     * Note: this can be extended in the future
     * @type {String}
     * @const
     */
    var AllowedPunctuatio = ":!\\?\\.,'`<>\\+\\-=/\\\\";

    /**
     * a list of punctuation marks that should be highligted seperately if
     * there is no sentence event
     * @type {String}
     * @const
     */
    var ReadablePunctuation = ":!\\?\\.,<>\\+\\-=/\\\\";

    AllowedPunctuatio = ":!\\?\\.,'`<>\\+\\-=/\\\\";
    ReadablePunctuation = "<>=/\\\\";

    /**
     * regexp to strip unreadable symbols
     * @type {RegExp}
     * @const
     */
    var SymbolStrip = new RegExp('[^'+AllowedPunctuatio+'\\w]', 'gi');

    /**
     * regexp for finding begin index
     * @type {RegExp}
     * @const
     */
    var BeginIndexRegEx = new RegExp('[\\w'+ReadablePunctuation+']', 'gi');

    /**
     * regexp for finding end index
     * @type {RegExp}
     * @const
     */
    var EndIndexRegEx = new RegExp('[\\s'+ReadablePunctuation+']', 'gi');

    function sanitize(text) {

        text = text.replace(/\u2013/g, '-');
        text = text.replace(/\u2014/g, '-');
        text = text.replace(/\u2019/g, '\'');
        text = text.replace(/\u2018/g, '\'');

        text = text.replace(/\n/g, ' ');
        text = text.replace(/\r/g, ' ');
        text = text.replace(/\t/g, ' ');

        // strip unreadable symbols
        text = text.replace(SymbolStrip, function(match, position) {
            return Array(match.length + 1).join(' ');
        }.bind(this));

        return text;
    }

    function getTextNodesIn(elem) {
      var textNodes = [];
      if (elem) {
        for (var nodes = elem.childNodes, i = 0; i < nodes.length; i++) {
          var node = nodes[i], nodeType = node.nodeType;
          if (nodeType == 3) {
              textNodes.push(node);
          } else if (nodeType == 1 || nodeType == 9 || nodeType == 11) {
            textNodes = textNodes.concat(getTextNodesIn(node));
          }
        }
      }
      return textNodes;
    }

    function processItem(node, selections) {
        var index = 0;
        var str = sanitize(node.data);
        var start;
        var end;
        var endReached = false;

        if($(node).closest('.mjx-chtml').length > 0) {
            return;
        }

        if($(node).closest('script').length > 0) {
            return;
        }

        while(true) {
            start = str.slice(index).search(BeginIndexRegEx);


            if(start == -1) {
                break;
            }

            start += index;

            if((new RegExp('['+ReadablePunctuation+']', 'gi')).test(str[start])) {
                end = start + 1;
            } else {
                end = str.slice(start).search(EndIndexRegEx);

                if(end == -1) {
                    end = str.length;
                    endReached = true;
                } else {
                    end += start;
                }
            }

            var range = document.createRange();
            range.setStart(node, start);
            range.setEnd(node, end);

            selections.push({
                node: node,
                range: range
            });

            if(endReached) {
                break;
            }

            index = end;
        }
    }

    function process(textNodes) {
        var selections = [];
        var item;

        for(var i = 0; i < textNodes.length; i++) {
            item = textNodes[i];

            processItem(item, selections);
        }

        return selections;
    }

    var nodes = getTextNodesIn($('.text').get(0));

    var selections = process(nodes);


    for(var i = 0; i < selections.length; i++) {
        (function(item, i) {
            setTimeout(function() {
                document.getSelection().removeAllRanges();
                var range = item.range;

                document.getSelection().addRange(range);
            }, i*500)
        }(selections[i], i))
    }
}());