By Mark Hanson
Has 1 other script.
/*
TODO: atom support / better feed format handling
TODO: choose one naming style, lower_case or xUpperCase (as inherited from RSS_Panel)
*/
// ==UserScript==
// @name keep_findory_current
// @namespace http://markerichanson.com/keep_findory_current
// @description Displays any articles findory hasn't found yet as if it had.
// @include
// @exclude
// ==/UserScript==
(function() {
// vars to handle data i need to store across async calls
var source_name = ""; // findory-used source name
var article_titles = [];
var script_rendered = []; // link->title for articles rendered by this script
// templates for HTML will insert later. probably a better way to save static
// strings in javascript.
var UNREAD_TEMPLATE = "\
<DIV class=\"article\">\
<DIV id=\"headline\">\
<A href=\"@link@\">@title@</A>\
</DIV>\
<DIV id=\"source\">\
<SPAN class=\"newforyou\">\
<A href=\"/source?source=@source_name@&ib=1\">@source_name@ </A>\
<EM>@when@ </EM>\
</SPAN>\
</DIV>\
<DIV id=\"blurb\" class=\"text\">@desc@\
<NOBR>\
<A href=\"@link@\"> (read more) </A> \
</NOBR> \
</DIV> \
</DIV>";
var READ_TEMPLATE = "\
<SPAN class=\"disabled\">\
<DIV class=\"article\">\
<DIV id=\"headline\">\
<A href=\"@link@\">@title@</A>\
</DIV>\
<DIV id=\"source\">\
<SPAN class=\"newforyou\">\
<A href=\"/source?source=@source_name@&ib=1\">@source_name@ </A>\
<EM>@when@ </EM>\
</SPAN>\
</DIV>\
<DIV id=\"blurb\" class=\"text\">@desc@\
<NOBR>\
<A href=\"@link@\"> (read more) </A> \
</NOBR> \
</DIV> \
</DIV>\
</SPAN>";
// begin the process of getting the feed xml for source on the current page
function rss_init() {
GM_log("rss_init()");
// get the rss url for the current source and pass it to get_and_handle_feed
yield_rss_url(get_and_handle_feed);
}
// get the rss url for the current source
// pass that url to the function argument
function yield_rss_url (rss_url_first_arg) {
GM_log("yield_rss_url("+source_name+")");
// have we stored the before?
rssUrl = GM_getValue(source_name);
if (rssUrl) {
GM_log(source_name+" stored rssUrl: "+rssUrl);
rss_url_first_arg (rssUrl);
}
else {
// helper function to save the url and pass it along to
// the target function.
function handle_url (url) {
GM_setValue(source_name,url);
rss_url_first_arg(url);
}
GM_log(source_name+" no stored rssUrl");
yield_xmlUrl_to_arg (
function (rssUrl) {
if (rssUrl == null) {
// not subscribed, so subscribe, extract, unsubscribe.
// oh what joy to do w/o a synchronous interface
yield_xml_url_by_adding_sub(handle_url);
}
else {
handle_url(rssUrl);
}
}
);
}
}
// 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) keep_findory_current'
},
onload: response_first_argument
});
}
// temporarily subscribe to the source in order to extract the xmlUrl
// from the resulting opml.
function yield_xml_url_by_adding_sub (rss_url_first_arg) {
GM_log("yield_xml_url_by_adding_sub("+source_name+",..)");
//TODO: really ought to parse the "add" & "remove" URLs from the button on page
// What the following does:
// 0. GET the "add favorite" URL for the source in question
// 1. call the function to retrieve the opml & extract the xmlUrl for
// the source in question.
// 2. pass the xmlUrl to an anonymous fuction that
// a. passes the xmlUrl to the original function argument
// b. removes the favorite by GETting the remove favorite url
// net net: added favorite, got xmlUrl, removed favorite.
// no other way to get the xmlUrl from Findory given the available data
// that I can see.
yield_get_response (
"http://findory.com/source?source="+source_name+"&ib=1&add_b=1",
function () {
yield_xmlUrl_to_arg(
function (xmlUrl) {
rss_url_first_arg(xmlUrl);
yield_get_response (
"http://findory.com/source?source="+source_name+"&ib=1&rem_b=1",
function () {
GM_log ("added, removed "+source_name+"as favorite.");
}
);
}
);
}
);
}
// get the feel url from the findory xml.
// pass the url to the function in the argument
function yield_xmlUrl_to_arg (xmlUrl_first_argument) {
GM_log("yield_xmlUrl_to_arg("+source_name+",..)");
// get the opml, craft a function on the fly to parse the opml
// and pass the results to argument xmlUrl_first_argument
yield_get_response (
"http://findory.com/export/opml/",
function (res) {
xmlUrl_first_argument(extract_xml_url_from_opml(res));
}
);
}
// parse the xmlUrl from the opml for the feed inidcated
function extract_xml_url_from_opml (res) {
GM_log("extract_xml_url_from_opml("+source_name+",..)");
query = '/opml/body/outline//outline[@title="'+source_name+'"]';
var dp = new XPCNativeWrapper(window, "DOMParser()");
var parser = new dp.DOMParser();
var DOM = parser.parseFromString(res.responseText, "application/xhtml+xml");
iter= document.evaluate(query,DOM, null, XPathResult.ANY_TYPE, null);
GM_log(iter);
i = iter.iterateNext();
if (i != null) {
rssUrl = null;
for (j = 0; (j < i.attributes.length) && (rssUrl == null); j++) {
if (i.attributes[j].nodeName == "xmlUrl") {
rssUrl = i.attributes[j].nodeValue;
}
}
//rssUrl = i ? (i.attributes["xmlUrl"]).nodeValue : null;
GM_log("rssUrl from current opml: "+rssUrl);
}
return rssUrl;
}
// get the feed xml indicated by the url and pass it to the handler
function get_and_handle_feed(src) {
if (! src) return;
/* TODO:intro caching for rss streams */
yield_get_response(src, handle_feed_response)
}
function handle_feed_response(res) {
// only if res is "loaded"
if (res.readyState == 4) {
// only if "OK"
if (res.status == 200) {
// handle result
// if-then-else inherited from RSS_Panel.
if (res.responseXML && typeof(res.responseXML) != "string") {
GM_log("0");
rss_render(res.responseXML);
// next try parsing the resultText with DOMParser;
// first the safe route through XPCNativeWrapper
}
else if (typeof(XPCNativeWrapper) == "function") {
GM_log("1");
var dp = new XPCNativeWrapper(window, "DOMParser()");
// dbg('XPC Wrapped DOM Parser: '+typeof(dp));
var parser = new dp.DOMParser();
// dbg('DOM Parser: '+typeof(parser));
var DOM = parser.parseFromString(res.responseText, "application/xhtml+xml");
rss_render(DOM);
// fallback to content window object; this would fail
// in GM 0.6.4+ but the safe option has succeeded already.
}
else if (typeof(window.DOMParser) == "object") {
GM_log("2");
// dbg("parsing DOMParser");
var parser = new win.DOMParser();
var DOM = parser.parseFromString(res.responseText, "application/xhtml+xml");
rss_render(DOM);
}
}
else {
dbg("GM_XHR response error: " + res.statusText);
}
}
}
function dom_getElements(node, elt) {
var list = node.getElementsByTagName(elt);
return (list.length) ? list : node.getElementsByTagNameNS("*", elt);
}
function dom_getFirstNodeValue(node, elt) {
try {
var list = dom_getElements(node, elt);
var chld = list[0].firstChild;
return chld.nodeValue;
} catch (e) {
// dbg("missing element " + elt + "\nError: " + e.message);
return "";
}
}
function normalize_title (title) {
var original_title = title;
title = title.toLowerCase();
title = title.replace(/'/g,'');
title = title.replace(/"/g,'');
// what was this for? some findory quirk with trailing spaces, but the
// semicolon may be a typo. yes, yes, I should have regression tests
// to catch this sort of thing, but this script started off as a quick
// small and dirty thing that wouldn't be maintained and i've not gone
// back to do tests. and, yes, minimally I should have commented the code.
title = title.replace(/\$[a-z]* ;/g,'');
// findory removes parenthesis from the title string as well
title = title.replace(/\([^}]*\)/g,'');
// pulled from http://www.aspdev.org/articles/javascript-trim/
// through the glory that is google. Surely repace(/\s*$/,'') or
// similar is better... and /^\s+/
while (title.substring(title.length-1, title.length) == ' ') {
title = title.substring(0,title.length-1);
}
while (title.substring(0,1) == ' ') {
title = title.substring(1, title.length);
}
// GM_log ("normalize_title ("+original_title+") = "+title);
return title;
}
function rss_render(DOM) {
findory_known_string = article_titles.join();
GM_log("known titles: "+findory_known_string);
var inner = "";
var items = [];
try {
items = dom_getElements(DOM, "item");
}
catch (e) {
}
unknown = 0;
for (var i=0; i<items.length; i++) {
// screen for already present, should be common case
var n = items[i];
var title = dom_getFirstNodeValue(n, "title") || "Untitled item #" + i;
var normalized_title = normalize_title(title);
// GM_log("feed title: "+normalized_title);
if (findory_known_string.indexOf(normalized_title) == -1) {
unknown++;
var desc = dom_getFirstNodeValue(n, "description");
desc = desc.replace(/<[^>]*>/g,"");
if (desc.length > 394) {
desc = desc.substring(0,394);
desc += "... ";
}
var link = (dom_getFirstNodeValue(n, "link") || "#RSS_MISSING_LINK");
var date = (dom_getFirstNodeValue(n, "date") || "NO DATE PARSED");
/* select template based on read/unread state saved */
var article = GM_getValue(source_name+normalized_title) == "read" ? READ_TEMPLATE : UNREAD_TEMPLATE;
article = article.replace(/@source_name@/g,source_name);
article = article.replace(/@link@/g,link);
article = article.replace(/@title@/g,title);
article = article.replace(/@when@/g,date);
article = article.replace(/@desc@/g,desc);
inner+=article;
script_rendered[link]=normalized_title;
// TODO: better sorting needed. when dates are available, could intermingle.
GM_log("article not known to findory:"+title+", "+desc.length);
}
}
//GM_log("unknown: "+unknown);
//GM_log(inner);
// insert what remains in the front of the articles section
query = '//TD[@id="slot_stripe_middle"]//div[@id="body"]';
iter = document.evaluate(query,document, null, XPathResult.ANY_TYPE, null);
// second id=body is the right one
var j = iter.iterateNext();
//var j = iter.iterateNext();
// var j = iter.iterateNext();
inner += j.innerHTML;
j.innerHTML = inner;
}
function onclick (event) {
normalized_title = script_rendered[event.target];
if (normalized_title != null) {
GM_log("script rendered article \""+normalized_title+"\" clicked.");
GM_setValue(source_name+normalized_title, "read");
}
else {
GM_log("findory rendered article clicked.");
}
GM_log(event.type+": "+event.target);
// if event.target is a link to one we rendered, save
// it for later as "read" for submission to Findory's
// history.
}
// pull all the interesting data off the page
function load_page_data () {
// get the title with the Findory : removed.
source_name = window.document.getElementsByTagName('title')[0].text.replace(/[^:]*: /,"");
query = '//div[@id="headline"]/a';
iter = document.evaluate(query,document, null, XPathResult.ANY_TYPE, null);
var j = iter.iterateNext();
while (j != null) {
var normalized_title = normalize_title(j.firstChild.nodeValue);
article_titles.push(normalized_title);
// if this article was read when it had been rendered by this script
// we need to tell findory it's read and change the display to mark
// it as read. Lastly, to prevent leaking into the GM storage, we
// need to remove the article's read flag.
if (GM_getValue(source_name+normalized_title) != null) {
// telling findory is easy. just GET the link.
yield_get_response(j.href, function () {});
// changing the display isn't too bad either. find the div containing
// this A, wrap it in a span with class "disabled" and reinsert the
// result as inner html
var theDiv = j.parentNode;
// have to call j's parent's parent's replaceChild
// method. That means we have to construct a replacement node
// it would be easier to change the innerHTML, but there's no node
// that is appropriate to do that too.
var newSpan = document.createElement("span");
newSpan.className="disabled";
newSpan.appendChild(theDiv);
theDiv.parentNode.replaceChild(newSpan,theDiv);5
// oddly, GM doesn't have a GM_removeValue. Hopefully this is close
// to the same net effect.
GM_storeValue(source_name+normalized_title, null);
}
j = iter.iterateNext();
}
}
// once the page loads, suck the data embedded on the page &
// kick off downloading the feed xml
function onload () {
// parse source title, item titles from page for later use
load_page_data();
rss_init();
}
function initialize () {
GM_log("initialize()");
// initialize
document.addEventListener("click", onclick, true);
window.addEventListener("load", onload, false);
}
initialize ();
})();
