Wayback Machine Small Bug Fixes

Fixes encoded ampersands on Wayback Machine's captures graph and problems that arise when trailing slashes are missing in an URL and other small issues universally present in all crawled sites
Verzia zo dňa 20.01.2016. Pozri najnovšiu verziu.
Nainštalovať tento skript ?
Spýtať sa otázku, uverejniť recenziu, alebo nahlásiť skript.
Zalomiť riadky
// ==UserScript==
// @name          Wayback Machine Small Bug Fixes
// @namespace     DoomTay
// @description   Fixes encoded ampersands on Wayback Machine's captures graph and problems that arise when trailing slashes are missing in an URL and other small issues universally present in all crawled sites
// @version       1.2.6
// @include       http://web.archive.org/web/*
// @include       http://wayback.archive.org/web/*
// @include       https://web.archive.org/web/*
// @include       https://wayback.archive.org/web/*
// @run-at        document-start
// @exclude       /\*/
// @grant         none

// ==/UserScript==

var toolbarNav = document.getElementById("wm-graph-anchor");
var lastFolder = window.location.href.substring(window.location.href.lastIndexOf("/") + 1);
var pics = document.images;
var backgrounds = document.querySelectorAll("[background]");
var shouldHaveTrailingSlash = (window.location.href.lastIndexOf(".") < window.location.href.lastIndexOf("/") || window.location.href.substring(window.location.href.lastIndexOf("//") + 2) == lastFolder) && lastFolder.indexOf("?") == -1;
var hasTrailingSlash = window.location.href.lastIndexOf("/") == window.location.href.length - 1;
var domain = window.location.href.substring(0,window.location.href.indexOf("/",window.location.href.lastIndexOf("//") + 2));

function fixToolbar()
{
	while(toolbarNav.href.indexOf("&amp;") > -1) toolbarNav.href = toolbarNav.href.replace("&amp;","&");
}

//Fix cases of &amp; in the capture graph
if(toolbarNav) fixToolbar();

if(!document.getElementsByTagName("base")[0])
{
	var base = document.createElement("base");
	if(shouldHaveTrailingSlash && !hasTrailingSlash) base.href = window.location.href + "/";
	else if((!hasTrailingSlash && !shouldHaveTrailingSlash) || hasTrailingSlash) base.href = document.baseURI;
	else base.href = domain + "/";
	document.head.appendChild(base);
}
	

for(var i = 0; i < pics.length; i++)
{
	//Skip over stuff related to the Wayback Machine toolbar and data URIs
	if((document.getElementById("wm-ipp") && document.getElementById("wm-ipp").contains(pics[i])) || pics[i].src.indexOf("data:") > -1) continue;
	//Refresh images in case the "base url" had to be modified.
	pics[i].src = pics[i].src;
	//For whatever reason, some images will point to within Internet Archive's "main" servers, instead of the crawled site. This attempts to fix that.
	if(pics[i].src.indexOf(document.domain + "/web") == -1) pics[i].src = domain.substring(0,domain.indexOf("/http") + 1) + pics[i].src.substring(pics[i].src.indexOf("/http") + 1);
}

for(var b = 0; b < backgrounds.length; b++)
{
	var bg = backgrounds[b].background || backgrounds[b].getAttribute("background");
	//Skip over stuff related to the Wayback Machine toolbar and data URIs
	if((document.getElementById("wm-ipp") && document.getElementById("wm-ipp").contains(backgrounds[b])) || bg.indexOf("data:") > -1) continue;
	//Refresh images in case the "base url" had to be modified.
	changeBackground(backgrounds[b],bg);
	//For whatever reason, some images will point to within Internet Archive's "main" servers, instead of the crawled site. This attempts to fix that.
	if(relativeToAbsolute(bg).indexOf(document.domain + "/web") == -1) 
	{
		var absoluteBG = relativeToAbsolute(bg)
		changeBackground(backgrounds[b],domain.substring(0,domain.indexOf("/http") + 1) + absoluteBG.substring(absoluteBG.indexOf("/http") + 1));
	}
}

function relativeToAbsolute(bgURL)
{
	var img = new Image();
	img.src = bgURL;
	return img.src;
}

function changeBackground(node, newBackground)
{
	if(node.background) node.background = newBackground;
	else if(backgrounds[b].getAttribute("background")) backgrounds[b].setAttribute("background",newBackground);
}

var observer = new MutationObserver(function(mutations) {
	mutations.forEach(function(mutation) {
		if(mutation.target.id == "wm-graph-anchor") toolbarNav = mutation.target;
		if(mutation.type == "attributes" && mutation.target == toolbarNav) fixToolbar();
		if(mutation.target.nodeName == "IMG" && mutation.attributeName == "src" && mutation.target.src.indexOf(document.domain + "/web") == -1)
		{
			mutation.target.src = domain.substring(0,domain.indexOf("/http") + 1) + mutation.target.src.substring(mutation.target.src.indexOf("/http") + 1);
		}
		if(mutation.attributeName == "background" && (mutation.target.getAttribute("background") || mutation.target.background).indexOf(document.domain + "/web") == -1)
		{
			var bg = mutation.target.background || mutation.target.getAttribute("background");
			var absoluteBG = relativeToAbsolute(bg);
			changeBackground(mutation.target,domain.substring(0,domain.indexOf("/http") + 1) + absoluteBG.substring(absoluteBG.indexOf("/http") + 1));
		}
	});    
});

var config = { attributes: true, childList: true, characterData: true, subtree: true };
observer.observe(document.body || document, config);