/*

Add blog post trackbacks to PubMed pages

*/

// ==UserScript==
// @name			Postgenomic / Pubmed mashup
// @namespace		http://www.postgenomic.com/userscripts
// @description		Adds blog post trackbacks to papers on PubMed
// @include			http://www.ncbi.nlm.nih.gov/entrez/*
// ==/UserScript==

// Get current date
var d = new Date();
var curr_date = d.getDate();
var stored_date=0;
var stored_date= GM_getValue("postgenomics_data_date", 0);

// Check for new DOIs in Postgenomic only once per day
// So if the current date != the date in the cache, refresh the cache
if ( (stored_date != curr_date) || (!GM_getValue("postgenomics_json_list",0)) ) {
	get_ids();
	GM_setValue("postgenomics_data_date", curr_date); // in JSON format
} else {
	eval(GM_getValue("postgenomics_json_list",0));
}

function process_api_results(obj) {
	GM_log("Got data, starting Postgenomic greasemonkey script.");
	
	var dois = new Array();
	var pmids = new Array();
	// ignore OAI ids for now
	
	dois = obj.doi_id; // an array of doi => paper_id
	pmids = obj.pubmed_id; // an array of PMID => paper_id	
	
	evaluate_page(dois, pmids);
}

function process_posts(obj) {
	var header = document.getElementById("GM_postgenomic");
	
	for (var i=0; i < obj.length; i++) {
		var title = obj[i].title;
		var blog_name = obj[i].blog_name;
		var pubdate = obj[i].pubdate;
		var link = obj[i].url;
		
		var post = document.createElement('li');
		post.innerHTML = '<a style="color: #1919B3;" href="' + link + '">' + title + '</a> in ' + blog_name;
		header.appendChild(post);
	}
}

function find(pattern, subject) {
	var reg = new RegExp(pattern, "i")
	var match = reg.exec(subject);
	return match; // returns an array (potentially empty)
}

function found_paper(paper_id) {
	var header = document.createElement('div');
	header.innerHTML = '<div id="GM_postgenomic" style="font-family: arial; font-size: 14px; background-color: #E6ECFF; color: #000000; border: 1px dotted #BFCFFF; padding: 5px; margin: 2px; margin-top: 15px;">' +
'<b>Trackbacks</b><br/>' +
'This paper has been cited on <a style="color: #1919B3;" href="http://www.postgenomic.com/">Postgenomic</a>.' +
'<p><br/></div>';

	var allPMIDs = document.evaluate(
		"//p[@class='pmid']",
	    document,
	    null,
	    XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE,
	    null);
	
	if (allPMIDs.snapshotLength) {
		var lastPMID = allPMIDs.snapshotItem((allPMIDs.snapshotLength - 1));
		lastPMID.appendChild(header);
	} else {
		// not sure where to put the script output... we'll just dump it at the top of the page.
		document.body.insertBefore(header, document.body.firstChild);		
	} 
	get_posts(paper_id);
		
}

function evaluate_page(dois, pmids) {
	GM_log("Evaluate page is running.");
	
	// first check the page URL
	var page_url = unescape(location.href);
	var match = find("list_uids=(\\d+)", page_url);
	if (match) {
		var pubmed_id = match[1];
		if (pmids[pubmed_id]) {
			found_paper(pmids[pubmed_id]);
			return;
		} else {
			GM_log("Found pubmed id " + pubmed_id + " but it isn't in the Postgenomic index.");
		}		
	} else {
		GM_log("Didn't find a match");
	}
	
	// Look for any PMIDs in the text
	GM_log("Looking for PMID in text");
	var allLinks, thisLink;
	var pmids_found = new Array();
	allLinks = document.evaluate(
		"//p[@class='pmid']",
	    document,
	    null,
	    XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE,
	    null);
	for (var i = 0; i < allLinks.snapshotLength; i++) {
	    thisLink = allLinks.snapshotItem(i);
	    // do something with thisLink
		GM_log("Found a PMID in text " + thisLink.textContent);
		var match = find("PMID: (\\d+)", thisLink.textContent);
		if (match) {
			var pubmed_id = match[1];
			pmids_found.push(pubmed_id);
		} else {
			GM_log("Found a PMID in text but it didn't match the regexp.");
		}
	}
	
	// we're only interested in pages where a single Pubmed ID is listed.
	if (pmids_found.length == 1) {
		var pubmed_id = pmids_found[0];
		if (pmids[pubmed_id]) {
			found_paper(pmids[pubmed_id]);
			return;
		} else {
			GM_log("Found single pubmed id " + pubmed_id + " in text but it isn't in the Postgenomic index.");
		}
	}
		
	return;
	
	var allLinks, thisLink;
	allLinks = document.evaluate(
		"//a[@href]",
	    document,
	    null,
	    XPathResult.UNORDERED_NODE_SNAPSHOT_TYPE,
	    null);
	for (var i = 0; i < allLinks.snapshotLength; i++) {
	    thisLink = allLinks.snapshotItem(i);
	    // do something with thisLink
		GM_log("Found a link " + thisLink.textContent);
	}	
}

// this function retrieves details of posts that cite a particular paper_id
function get_posts(paper_id){
	GM_log("Requesting posts");
	var request_url = "http://www.postgenomic.com/api.php?type=post&citing_paper=" + paper_id + "&format=json&callback=process_posts&about=fnl_pubmed_trackbacks";
	GM_log(request_url);
	GM_xmlhttpRequest({
		method: 'GET',
		url: request_url,
		headers: {
			'User-agent': 'Mozilla/4.0 (compatible) Greasemonkey',
			'Accept': 'application/xml,text/html',
		},
		onload: function(responseDetails) {
			var response_status=responseDetails.status;
			var response_text=null;
			if (response_status==200){
				GM_log("Posts request: response ok");
				response_text=responseDetails.responseText;
				eval(response_text);
			}
		}
	});
}

// this function retrieves details of papers in the Postgenomic index in JSON format.
function get_ids(){
	GM_log("Requesting IDs");
	GM_xmlhttpRequest({
		method: 'GET',
		url: "http://www.postgenomic.com/api.php?type=paper&ids_only=1&callback=process_api_results&about=fnl_pubmed_trackbacks",
		headers: {
			'User-agent': 'Mozilla/4.0 (compatible) Greasemonkey',
			'Accept': 'application/xml,text/html',
		},
		onload: function(responseDetails) {
			var response_status=responseDetails.status;
			var response_text=null;
			if (response_status==200){
				GM_log("IDs request: response ok");
				response_text=responseDetails.responseText;
			
				GM_setValue("postgenomics_json_list", response_text);
				eval(response_text);
			}
		}
	});
}