rojo_del.icio.us
By Mark Hanson
—
Last update Feb 10, 2006
—
Installed
613 times.
// Rojo Del.icio.us
// version 0.1 Alpha
// 2006-02-02
// Copyright (c) 2006, Mark Hanson
// Released under the GPL license
// http://www.gnu.org/copyleft/gpl.html
// ==UserScript==
// @name rojo_del.icio.us
// @version 0.1
// @description Post tags applied in Rojo as bookmarks in del.icio.us
// @namespace http://markerichanson.com/greasemonkey/rojo
// @include http://www.rojo.com/*
// @include http://rojo.com/*
// @exclude http://www.rojo.com/javascript/dojo/iframe_history.html
// @exclude http://rojo.com/javascript/dojo/iframe_history.html
// Whats New
// =========
// v 0.1 First cut
//
// Description
// ===========
// Tag an arcticle of feed in rojo, and it's tagged in del.icio.us. Delete a tag
// in rojo and it's deleted in del.icio.us. Why? Because. Plus, it gave me an
// excuse to play with DOM mutation events.
//
// To-Do:
// Handle tags on feed management page. Started, not working. See the to-be-renamed
// getPostItem2...
// Abstract the del.icio.us portion to allow plug in of other targets.
// Reflect changes made at del.icio.us to tags into rojo.
// Contact
// =======
// Mark Hanson ( http://markerichanson.com/ )
// ==/UserScript==
(function() {
var h = {};
function initialize () {
// GM_log("initialize("+unsafeWindow.location+")");
// GM_log("initialize("+unsafeWindow.location.pathname+")");
// initialize
// initialize page specific handler array
h["/doIRPCTagArticle.jsp"] = doPostItem;
h["/doIRPCdeleteTagArticle.jsp"] = doPostItem;
h["/doIRPCTagFeed.jsp"] = startTagFeed;
h["/view-feed/"] = endTagFeed;
h["/doIRPCdeleteTagFeed.jsp"] = startTagFeed;
h["/manage-feeds/doIRPCTagFeed.jsp"] = asYetUnnamed;
// defer activities until load event fires.
window.addEventListener("load", onLoad, false);
}
initialize ();
// looking for events to identify
// assuming some special processing will be required on the add/delete tag
// pages for the /manage-feeds/ handling
function asYetUnnamed () {
}
// managing tag feeds takes two steps. First, here, note the GET to the
// path that causes tags on feeds to be added or deleted (via "h[]" lookup
// on load. On this event, pull the records stored with GM_setValue and
// load an ifrane with the info url for the feed in rojo. Wait for the
// onload call of endTagFeed for that iframe.
// btw, I'd have done this with a GM_xmlhttprequest and domparser, but rojo's
// html wasn't well-formed when I was writing this so the domparser bonked.
function startTagFeed () {
// have to get the rojo feed description to get the real url before
// posting.
// parse out //div[@class='description']/A[1]
GM_log("startTagFeed");
var src = GM_getValue("postItem");
var o = eval(src);
// some of this iframe code came from:
// http://youngpup.net/userscripts/htmlparserexample.user.js
// but the approach didn't. Using the greasemonkey hooks to do onload
// processing for the page in the iframe isn't how the link does it.
// doing it this way works b/c i'm passing data between instances of the
// script with GM_setValue & GM_getValue which is, no doubt, an abuse of
// some rule somewhere..
// create an IFRAME to write the document into. the iframe must be added
// to the document and rendered (eg display != none) to be property
// initialized.
var iframe = document.createElement("IFRAME");
iframe.style.visibility = "hidden";
iframe.style.position = "absolute";
document.body.appendChild(iframe);
// set the iframe to load the link from the PostItem. This loads the
// rojo feed info display from which I can parse (in endTagFeed) the
// sites URL which is otherwise not available on the page where the tag
// was applied.
iframe.contentWindow.location.href = o.url;
}
// called when the feed info page is loaded. Parses the feed URL from that
// page, updates the PostItem stored in GM_setValue ("postItem"...). Calls
// doPostItem so that the feed tag is reflected into del.icio.us.
function endTagFeed () {
GM_log("endTagFeed");
var src = GM_getValue("postItem");
var o = eval(src);
var query = "//div[@class='description']/A[1]";
var iter= document.evaluate(query,document, null, XPathResult.ANY_TYPE, null);
var i = iter.iterateNext();
GM_log (i+": "+i.href);
o.url = i.href;
GM_setValue("postItem",o.toSource());
doPostItem();
}
// parse tag text out of node. trim.
function getTag (node) {
// GM_log("getTag: "+node);
// text from the first <a> child is the tag
var query = "./a[1]/text()";
iter= document.evaluate(query,node,null, XPathResult.ANY_TYPE, null);
i = iter.iterateNext();
return i.textContent.replace(/\n/g,'').replace(/^\s+/g, '').replace(/\s+$/g, '');
}
// constructor for tuple that is a rojo tag / del.icio.us post
function PostItem (url, title, tags) {
this.url = url;
this.title = title;
this.tags = tags;
}
// parse a post item off of the /manage-feed/ html.
// partial implementation, untested xpath. not working currently.
function getPostItem2 (node) {
var query = "..//li";
var iter= document.evaluate(query,node,null, XPathResult.ANY_TYPE, null);
li = iter.iterateNext();
var tags = new Array();
var url = "";
var title = "";
while (li != null) {
tags.push(getTag(li));
li = iter.iterateNext();
}
GM_log("tags: "+tags.toSource());
query="../../../TD[class='subscription-info']/A";
iter = document.evaluate(query,document,null, XPathResult.ANY_TYPE, null);
url = iter.iterateNext().href;
GM_log("url: "+url);
query="../../../TD[class='subscription-title']/SPAN";
iter = document.evaluate(query,document,null, XPathResult.ANY_TYPE, null);
title = iter.iterateNext().textContent;
GM_log("title: "+title);
// must be the manage feed page.
// title
// link is the [i] image
return new PostItem(url,title, tags);
}
// parse post items out of html.
// a little ugly: try to parse as if it's part of an article tag first. if
// that fails, then parse as a feed tag.
// check at the outset if it's on /manage-feeds/ and send out to the other
// method. Doing it that way is better than forking a separate onDOMNodeInserted
// listener and registration based on the current location.
function getPostItem (node) {
// ../../../span has the link if an article,
// ../../../../../../H1 has the rojo link and title if feed
// distinguish - if there's a span sibling to ../../.., then article, else
// feed
var path = unsafeWindow.location.pathname;
if (path == "/manage-feeds/") {
return getPostItem2(node);
}
// need to ignore LI without IDs. looks like a rendering bug on
// rojo's side that ghost tags are showing up without ids.
var query = "..//li[@id]";
var iter= document.evaluate(query,node,null, XPathResult.ANY_TYPE, null);
li = iter.iterateNext();
var tags = new Array();
while (li != null) {
tags.push(getTag(li));
li = iter.iterateNext();
}
query="../../../../span[1]";
iter= document.evaluate(query,node,null, XPathResult.ANY_TYPE, null);
i = iter.iterateNext();
var title = "";
var url = "";
if (i != null) {
// it's an article being tagged
// 12 hours later, what the hell was the line below? what i get for
// coding between playing with a 6 year old and 3 year old...
// seems to work, oddly enough.
url = i.title.replace(/\n/g,'').replace(/^\s+/g, '').replace(/\s+$/g, '');
query="./a/text()"
iter = document.evaluate(query,i,null, XPathResult.ANY_TYPE, null);
title = iter.iterateNext().textContent;
title = title.replace(/\n/g,'').replace(/^\s+/g, '').replace(/\s+$/g, '');
}
else {
// must be the feed itself
query="//H1/text()";
iter = document.evaluate(query,document,null, XPathResult.ANY_TYPE, null);
var titleElement = iter.iterateNext();
if (titleElement != null) {
title = titleElement.textContent;
query="//H1/SPAN/A[2]";
iter = document.evaluate(query,document,null, XPathResult.ANY_TYPE, null);
url = iter.iterateNext().href;
}
}
return new PostItem(url,title, tags);
}
// note that a tag is removed. create and store a PostItem.
// posting will be done when the page rojo uses for back end communication
// loads.
function onDOMNodeRemoved (event) {
GM_log("onDOMNodeRemoved: "+event.target);
GM_log("event.target.className: "+ event.target.className);
if (event.target.className == "tagLi") {
// tags removed are not yet removed from the postItem
// remove it here.
var theTag = getTag(event.target);
var postItem = getPostItem(event.target);
var newTags = new Array();
// remove tag from tags
for (var i = 0; i < postItem.tags.length; i++) {
if (postItem.tags[i] != theTag) {
newTags.push(postItem.tags[i]);
}
}
postItem.tags = newTags;
GM_log("postItem: "+postItem.toSource());
GM_setValue("postItem",postItem.toSource());
}
}
// note addition of a tag. Create a post item and store via gm_setvalue.
// if multiple tags are added, a PostItem will be created for each and
// only the final one left stored via gm_setvalue (all but last are overwritten)
// actual post of postitem will happen when the page rojo uses for back end
// communication is written.
function onDOMNodeInserted (event) {
GM_log("onDOMNodeInserted: "+event.target);
GM_log("event.target.className: "+ event.target.className);
if (event.target.className == "tagLi") {
// var theTag = getTag(event.target);
// tags added are already in the postItem.
// no need to add it here.
var postItem = getPostItem(event.target);
GM_log("postItem: "+postItem.toSource());
GM_setValue("postItem",postItem.toSource());
}
}
// Post whatever post item is currently stored. If the currently stored item
// has 0 tags, treat it as a delete.
// NOTE: this relies on you being logged into del.icio.us. You will be
// prompted to login if you are not. Not sure what happens when you log out
// of del.icio.us. Early testing indicated posting would somehow still work
// but I'm not sure...
function doPostItem () {
GM_log("doPostItem");
var src = GM_getValue("postItem");
GM_log("src: "+src);
// if tags.length == 0, then delete
var o = eval(src);
if (o.tags.length == 0) {
// delete
var url = "http://del.icio.us/api/posts/delete?";
url += "url="+o.url;
yield_get_response(url, function (resp) {
GM_log("response received: "+resp.toSource());
});
}
else {
var url = "http://del.icio.us/api/posts/add?";
url += "&tags="+o.tags.join(" ");
url += "&description="+o.title;
url += "&url="+o.url;
yield_get_response(url, function (resp) {
GM_log("response received: "+resp.toSource());
});
}
}
// Issue a GET to the specified URL, then pass the response to the
// function argument. Utility method.
// TODO: make this degrade gracefully?
function yield_get_response (url, response_first_argument) {
GM_log("yield_get_response("+url+",..)");
GM_xmlhttpRequest ({
method: 'GET',
url: url,
headers: {
'User-agent': 'Mozilla/4.0 (compatible) rojo_delicios.user.js'
},
onload: response_first_argument
});
}
function onLoad () {
// run the page-specific function, if any.
var path = unsafeWindow.location.pathname;
var f = h[path];
if (f != null) f();
// put the mutation event handlers in place if this is a window with
// TagArticle available.
// in the end, the tags might be captured more easily by
// intercepting specific page logs (doIRPCTagArticle etc) but the
// mutation events were fun to play with so I'm leaving them in until
// there's some compelling reason to rewrite their functionality.
// later note: not so sure it would be doable without mutation events.
// mutation events are in-place in the html and i can get to the rest of
// the tags on an item. at least some of the doIRPC... methods on pass
// the changed tag & del.icio.us api doesn't allow passing only the
// changed tags.
// only put mutation event listeners on pages that have access to the
// javascript object that makes the mutations happen. this is an attempt
// to avoid monkeying with all the iframe-based backend communication
// rojo does.
if (unsafeWindow.TagArticle != null) {
// document.addEventListener("click", onEvent, false);
// document.addEventListener("DOMAttrModified", onEvent, false);
// document.addEventListener("DOMCharacterDataModified", onEvent, false);
document.addEventListener("DOMNodeInserted", onDOMNodeInserted, false);
// document.addEventListener("DOMNodeInsertedIntoDocument", onEvent, false);
document.addEventListener("DOMNodeRemoved", onDOMNodeRemoved, false);
// document.addEventListener("DOMNodeRemovedFromDocument", onEvent, false);
// document.addEventListener("DOMSubtreeModified", onEvent, false);
}
}
})();
