Scrape(General|Single): Get Paragraphs & Sentences

//this method focuses on extracting the chunks of text from nodes that most likely contain sentences or paragraphs.
[...[...document.querySelectorAll("*")].reduce((s, el) => {
	//look for text that meets certain credentials
	[...el.childNodes].forEach(child => {
	child.nodeName == "#text" && child.nodeValue.replace(/[\s\t\r\n]+/gim," ").trim().length > 1 ? s.add(child.nodeValue.replace(/[\n\r]+/gim,"\n").replace(/[ \t]+/gim," ").trim()) : child.innerText && child.innerText.replace(/[\s\t\r\n]+/gim," ").trim().length > 80 && child.innerText.replace(/[\s\t\r\n]+/gim," ").trim().length < 3000 ? s.add(child.innerText.replace(/[\n\r]+/gim,"\n").replace(/[ \t]+/gim," ").trim()) : null;
  })
return s;
},new Set())].filter(str => {
//second round of filtering...

//count # of new lines in the string. When line count > punct count, assume it's not a sentence / paragraph
var newlineCount = str.match(/\n/gim);
		newlineCount = newlineCount ? newlineCount.length : 0 ;
//count punctuation so it can be compared to newlinecount
var punctCount = str.match(/[.?!]/gim);
		punctCount = punctCount ? punctCount.length : 0 ;
//count spaces (should only be spaces between words due to previous cleanup process)
var spaceCount = str.match(/[ ]/gim);
		spaceCount = spaceCount ? spaceCount.length : 0 ;
//estimate word count by finding word boundaries next to alphanum chars
var estWordCount = str.match(/[a-z0-9]\b/gim);
		estWordCount = estWordCount ? estWordCount.length : 0 ;
//only accept text that starts with a number or capped letter
var rx = RegExp('^[A-Z0-9]', 'g');
var startsAlphaNum = rx.test(str);
if(!startsAlphaNum || estWordCount <= 3){
	return false;
}else if(newlineCount > 0){

var ratio = punctCount/newlineCount;
//low ratio = probably not a sentence/paragraph... keep only strings with high punct to newline ratios
return ratio > 0.75;
}else{
	return true;
}

});