articleParser.js

(function(){
	var script = document.createElement('script');
	script.type = 'text/javascript';
	script.src = "https://ajax.googleapis.com/ajax/libs/jquery/3.3.1/jquery.min.js"
	document.head.appendChild(script);

	
	var imageCounter = 0;
	
	function getLinkFromHeader(item){
		var a = $(item).find('a')[0];
		if(a && a.href){
			return a.href;
		}	
	}
	
	var handlers = {
		H1: function(item){
			var text = item.textContent;
			var link = getLinkFromHeader(item);
			if(link){
				return '# ' + '[' + text + ']' + '(' + link + ')';
			}
			return '# ' + text;
		},
		H3: function(item){
			var text = item.textContent;
			var link = getLinkFromHeader(item);
			if(link){
				return '## ' + '[' + text + ']' + '(' + link + ')';
			}
			return '## ' + text;
		},
		H4: function(item){
			var text = item.textContent;
			var link = getLinkFromHeader(item);
			if(link){
				return '### ' + '[' + text + ']' + '(' + link + ')';
			}
			return '### ' + text;
		},
		IMG: function(item){
			console.log(item.src);
			var link = item.src;
			imageCounter++;
			return '![' + imageCounter + '](' + link + ')';
		},
		A: function(item){
			var link = item.href,
				text = item.textContent;
			return '[' + text + '](' + link + ')';
		},
		IFRAME: function(item){
			var scripts = $($(item).contents()[0]).find('script');
			if(scripts && scripts.length && scripts[0]){
				var link = scripts[0].src;
				if(link.indexOf('git')){
					return '<script src=' + link +'></script>'
				} return undefined;
			} return undefined;
		},
		PRE: function(item){
			return '```' + item.innerText + '```';
		},
		P: function(item){
			return '\r\n' + item.innerText + '\r\n';
		},
		BLOCKQUOTE: function(item){
			return '\r\n' + item.innerText + '\r\n';
		},
		LI: function(item){
			return '\r\n * ' + item.innerText + '\r\n';
		}
	}
	
	var items = $('p, blockquote, .graf.graf--h3, .graf.graf--h4, img.progressiveMedia-image.js-progressiveMedia-image, a.markup--anchor.markup--p-anchor, li.graf--li, div.iframeContainer iframe, .graf.graf--pre.graf-after--p');
	window.resultItems = [];
	$.each(items, function(index, item){
		var processFunc = handlers[item.tagName.toUpperCase()];
		if(processFunc){
			var result = processFunc(item);
			window.resultItems.push(result);
		}
	})
	
	console.log(window.resultItems.join('\r\n\r\n'));
})()

(function(){
	var shift = 0;
	var handlers = {
		H1: function(item, prev){
			var text = item.textContent;
			return prev + '\r\n' + '# ' + text;
		},
		H3: function(item, prev){
			var text = item.textContent;
			return prev + '\r\n' + '## ' + text;
		},
		H4: function(item, prev){
			var text = item.textContent;
			return prev + '\r\n' + '### ' + text;
		},
		PRE: function(item, prev){
			return prev + '\r\n' + '```' + item.innerText + '```';
		},
		P: function(item, prev){
			console.log(prev);
			return prev + '\r\n' + item.innerText + '\r\n';
		},
		BLOCKQUOTE: function(item, prev){
			return prev + '\r\n' + item.innerText + '\r\n';
		},
		LI: function(item, prev){
			return prev + '\r\n * ' + item.innerText + '\r\n';
		}
	}
	
	var items = $('p, blockquote, .graf.graf--h3, .graf.graf--h4, img.progressiveMedia-image.js-progressiveMedia-image, a.markup--anchor.markup--p-anchor, li.graf--li, div.iframeContainer iframe, .graf.graf--pre.graf-after--p');
	$.each(items, function(index, item){
		var tagName = item.tagName.toUpperCase();
		var processFunc = handlers[tagName];
		if(tagName === 'A'){ //wrong translation of A element (duplicate)
			var link = item.href;
			if(window.resultItems[index-shift-1].indexOf(link) >-1) {
				shift+= 1;
			}
		}
		if(processFunc){
			window.resultItems[index - shift] = processFunc(item, window.resultItems[index - shift]);
		}
		
	})
	console.log(window.resultItems.join('\r\n\r\n'));
})()