Scrape(General|Single|NLP): Get Keyword Stats - Snippet Editor

Save Snippet View Snippet

(async () => {
  var options = {
    scripts: ["https://d3js.org/d3.v6.min.js"],
  }
  const downloadCSV = (data, fileName) => {
		var d = document;
    var a = d.createElement('a');
    a.href = window.URL.createObjectURL(new Blob([d3.csvFormat(data)],{type:'text/csv;charset=utf-8;'}));
    a.setAttribute('download', `${fileName}.csv`);
    d.body.appendChild(a);
    a.click();
    d.body.removeChild(a);
  };
	var run = async () => {
 var stripTags = ["SCRIPT","STYLE","NOSCRIPT"];
  var stopWords = [ "by","of","to","as","at","an","or","on","no","be","from","with","more","for","have","want", "than","and","can","the","are","not","arent","aren't","i", "we", "he", "she", "it", "they", "me", "us", "her", "him", "them", "mine", "ours", "hers", "his", "theirs", "my", "our", "your", "their", "myself", "yourself", "herself", "himself", "itself", "ourselves", "yourselves", "themselves", "all", "another", "any", "anybody", "anyone", "anything", "both", "each", "either", "everybody", "everyone", "everything", "few", "many", "most", "neither", "nobody", "none", "no one", "nothing", "one", "other", "others", "several", "some", "somebody", "someone", "something", "such", "that", "these", "this", "those", "what", "whatever", "which", "whichever", "who", "whoever", "whom", "whomever", "whose", "as", "thou", "thee", "thy", "thine", "ye", "aught", "each other", "enough", "idem", "its", "naught", "nought", "one another", "ought", "ourself", "somewhat", "suchlike", "theirself", "theirselves", "themself", "there", "thyself", "whatnot", "whatsoever", "whence", "where", "whereby", "wherefrom", "wherein", "whereinto", "whereof", "whereon", "wherever", "wheresoever", "whereto", "whereunto", "wherewith", "wherewithal", "whether", "whichsoever", "whomso", "whomsoever", "whosever", "whosesoever", "whoso", "whosoever", "yon", "yonder", "you", "yours" ];
var extractTextFromTag = (tag) => {
     // console.log("tag",tag)
			try{
				tag.textContent = ""+tag.textContent.replace(/[^a-z0-9\s'-]/gim, " ").replace(/[\s\t\r\n]+/gim, " ").trim();
				return ((tag.textContent.length > 0) && tag.nodeName == "#text") ? tag.textContent : null;
			}catch(err){
 				console.log(err,tag);
      }
	}
	var extractTextFromTags = (tags) => {
		return tags.reduce((result,tag) => {
		var tagText = extractTextFromTag(tag);
		(tagText) ? result.push(tagText): null;
		return result;},[]);
  }
	var processChildren = (parent) => {
		return [...parent.childNodes].reduce((result,tag) => {(stripTags.indexOf(tag.tagName) == -1) ? result.push(tag) : null; return result;},[]);
  }
	var children = await processChildren(document);
  var childrenWithNoChild = [];
  for (var i = 0; i < children.length; i++) {
		var child = children[i];
    if(stripTags.indexOf(child.tagName) == -1)
    (children[i].hasChildNodes()) ? children = children.concat([...child.childNodes]): childrenWithNoChild.push(child);
  }

downloadCSV(keysSorted, "wordcount_"+document.location.hostname.replace(".","_"))
  }
  const loadScripts = ({scripts}) => {
    console.log(`Loading External Scripts`);
    var scriptCountdown = scripts.length;
    var loadScript = (url) => {
      var scriptsLoaded = () => scriptCountdown == 0 ? run() : null;
      var imported = document.createElement('script');
      imported.src = url;
      imported.addEventListener("load", () => {
        scriptCountdown--;
        scriptsLoaded();
      });
      document.head.appendChild(imported);
    }
    scripts.forEach(loadScript)
  }
  loadScripts(options);

})();