By pluto2
—
Last update
Nov 26, 2007
—
Installed
2,335 times.
Add Syntax Highlighting (this will take a few seconds, probably freezing your browser while it works)
// PubMed Citations v0.31
// version 0.3
// 11-26-07
// Copyright (c) 2007, Pluto
// email: pluto@plutosforge.com
// ==UserScript==
// @name PubMed Citations
// @description lists the number of citations(with a link) next to each search result in pubmed
// @namespace userscripts.org/scripts/show/13704
// @include http://www.ncbi.nlm.nih.gov/sites/entrez
// ==/UserScript==
var thisVersion = '0.31'; //need this for checking for updates at userscripts.org
var greekReSearchNotice = ' (Too much Greek?, re-searching...)';
var badGreekTitles = 0; //an index for the badTitleResultLines array
var badTitleResultLines = new Array(); //an array to store the resultLines of the greek titles to re-search
var authorReSearchNotice = ' (No results found, trying author search...)';
var authorReSearches = 0;
var authorReSearchLines = new Array();
var allResults = new Array(); //an array to associate the <div> numbers of a result with it's title, and other possible info
var allResultsIndex = 0; //an indexer to keep track of the total number of titles found
//** need to put the variable declarations first because I guess greasemonkey does not instantiate them all first whereever they are, which is a problem for slower machines
//will wait a total of 5 sec, but check each 0.25 sec to see if the previous set of searches has finished yet
const maxWaits = 20;
const waitTime = 250;
var numWaits= 1; //keeps track of the number of times the recursive function has been called
var timeOutCounter = 0; //for debugging purposes - keeps track of the total time elapsed
var firstHaveAllComments = false;
main(); //DO IT!
function main()
{
//in Pubmed's new design (11-9-07) each results title gets its own <div> tag
//so this goes through all the <div> tags, finds the ones named "title", then pulls out the title, formats it and sends it off to Google as before
for (var i = 0; i < document.getElementsByTagName('div').length; i++)
{
if (document.getElementsByTagName('div')[i].className == 'title')
{
getCitations(getTitle(i),i,'first');
//store each title in an array for use later
allResults[i] = getTitle(i);
allResultsIndex++;
}
}
firstRecursiveWait(); //wait for all the title searches to be done, then research the ones with greek titles that didn't return results
}
function firstRecursiveWait()
{
GM_log('in first wait' + waitTime + '__' + numWaits);
window.setTimeout(function()
{
//if havent found all the comments yet (i.e. done searching) and havent maxed out the counter yet, then go on; else will be empty settimeout function
if(!firstHaveAllComments && numWaits < maxWaits)
{
//increment the counters before the recursive call
timeOutCounter += waitTime;
numWaits++;
//check to see if all the results have had comments added (i.e. the search is done)
if(checkResultComments() == allResultsIndex)
{
//if done, then set the trip variable
haveAllComments = true;
GM_log(timeOutCounter + ' ' + 'finally done' + ' ' + checkResultComments() + ' =? ' + allResultsIndex);
//startup the google search for each specified result
for (var i = 0; i < badGreekTitles; i++)
{
var badTitle = getTitle(badTitleResultLines[i]);
//change the spelled out greek characters to the actual letters - the whole point of this re-searc
badTitle = replaceGreek(badTitle);
//send 'greek' as the third argument to tell getCitations() that this is a re-search for a greek title
getCitations(badTitle,badTitleResultLines[i],'greek');
}
//now, reset the global counters before calling the second recursive call function
numWaits= 1;
timeOutCounter = 0;
secondRecursiveWait(); //check for all the greek searches to be done, then call the authors search
}
else //then still waiting, so call recursive call again
{
GM_log(timeOutCounter + ' ' + 'not done yet' + ' ' + checkResultComments() + ' =? ' + allResultsIndex);
firstRecursiveWait();
}
}
}, waitTime * numWaits); //sets the delay longer and longer as the recursion advances
}
//have a second loop trip because resetting the original one will cause the previous function to restart searching for comments
var secondHaveAllComments = false;
function secondRecursiveWait() //this function is nearly identical to the first one
{
window.setTimeout(function()
{
if(!secondHaveAllComments && numWaits < maxWaits)
{
timeOutCounter += waitTime;
numWaits++;
if(checkGreekReSearchComments() == badTitleResultLines.length)
{
//debug GM_log(timeOutCounter + ' ' + '2nd finally done' + ' ' + checkGreekReSearchComments());
secondHaveAllComments = true;
for (var i = 0; i < authorReSearches; i++)
{
//here the difference is that we search with authors instead of titles
var authors = getAuthors(authorReSearchLines[i]);
//debug GM_log('about to search for: ' + authors);
//send 'authors' as the third argument to tell getCitations() that this is a re-search with just authors
getCitations(authors,authorReSearchLines[i],'author');
checkForMultipleScripts();
}
checkForUpdates();
}
else
{
//debug GM_log(timeOutCounter + ' ' + '2nd not done yet' + ' ' + checkGreekReSearchComments());
secondRecursiveWait();
}
}
}, waitTime * numWaits);
}
//######################################## ##############################################
//################################## FORMATTING FUNCTIONS ########################################
//####################################### ##############################################
//each HTML comment has "!" as part of it, so lthis funciton looks for that to denote a comment and proof that the search finished
function checkResultComments()
{
var numComments = 0;
for(var i in allResults)
{
if(document.getElementsByTagName('div')[parseInt(i)+2].innerHTML.indexOf("!") != -1)
{
numComments++;
}
}
return numComments;
}
//after a greek title re-search, I insert a "^greek search done", so lthis funciton looks for the "^" to denote show that the search finished
function checkGreekReSearchComments()
{
var numComments = 0;
for(var i = 0; i < badTitleResultLines.length; i++)
{
if(document.getElementsByTagName('div')[badTitleResultLines[i]+2].innerHTML.indexOf("^") != -1)
{
numComments++;
}
}
return numComments;
}
//pull the title from the page and replace all the spaces with +'s for the search query ---- may not have to do this - google may do it automatically - someting to try out
function getTitle(resultLine)
{
var title = document.getElementsByTagName('div')[resultLine].textContent;
while ((title.indexOf(" ") != -1) || (title.indexOf("&") != -1))
{
title = title.replace(' ','+');
title = title.replace('&','%26'); //ampersands screw up the URL
}
return title;
}
//pull the authors from a result and remove the first,middle initials and just make a list of the last names because that's fine with google... actually initials last give no results
function getAuthors(resultLine)
{
//this pops up, and back in through the CSS to get the list of authors... good chance this will be the first to break with a Pubmed page redesign
var authorsString = document.getElementsByTagName('div')[resultLine].parentNode.parentNode.previousSibling.previousSibling.childNodes[1].textContent;
//put all the authors into an array using ", " as the delimiter
var authorsArray = authorsString.split(', ');
var lastNames = ''; //a string of the author last names, separated by commas
for(var i = 0; i < authorsArray.length; i++)
{
//this next line will take out the last name from the array, leaving the space that was separating the last name and initials
lastNames += authorsArray[i].substring(0,authorsArray[i].indexOf(' '));
if(i < authorsArray.length-1)
{
//don't slap on the comma if you're on the last author
lastNames += ', ';
}
}
return lastNames;
}
//replace all the spelled out greek words with the actual symbols
function replaceGreek(thisTitle)
{
//set these string variables to the ASCII codes
var alpha = String.fromCharCode(945);
var beta = String.fromCharCode(946);
var gamma = String.fromCharCode(947);
var delta = String.fromCharCode(948);
var kappa = String.fromCharCode(954);
var mu = String.fromCharCode(956);
while ((thisTitle.indexOf('alpha') != -1) ||
(thisTitle.indexOf('beta') != -1) ||
(thisTitle.indexOf('gamma') != -1) ||
(thisTitle.indexOf('delta') != -1) ||
(thisTitle.indexOf('kappa') != -1) ||
(thisTitle.indexOf('mu-') != -1) || (thisTitle.indexOf('mu ') != -1) )
{
thisTitle = thisTitle.replace('alpha',alpha);
thisTitle = thisTitle.replace('beta',beta);
thisTitle = thisTitle.replace('gamma',gamma);
thisTitle = thisTitle.replace('delta',delta);
thisTitle = thisTitle.replace('kappa',kappa);
thisTitle = thisTitle.replace('mu-',mu + '-'); // only deal with these two cases to not screw up words like muscarinc
thisTitle = thisTitle.replace('mu ',mu + ' ');
}
return thisTitle;
}
//######################################## ##############################################
//################################## SEARCHING FUNCTIONS ########################################
//####################################### ##############################################
function getCitations(thisTitle,resultLine,searchMode)
{
var searchURL;
if(searchMode == 'author')
{
//then don't need the "intitle:" prefix
searchURL = 'http://scholar.google.com/scholar?hl=en&lr=&q=' + thisTitle + '&btnG=Search';
}
else // add in the intitle prefix
{
searchURL = 'http://scholar.google.com/scholar?hl=en&lr=&q=intitle%3A%22' + thisTitle + '%22&btnG=Search';
}
GM_xmlhttpRequest(
{
method: 'GET',
url: searchURL,
headers: {/* for some odd reason sending these headers caused Google to return a search with a greek character in the title as have no results when a manual search in the browser showed that there clearly was a result
'User-agent': 'Mozilla/4.0 (compatible) Greasemonkey/0.3', //so, commenting everything out sends the default blank string header and it works fine now
'Accept': 'application/atom+xml,application/xml,text/xml',*/},
onload: function(responseDetails)
{
//an object for storing and manipulating the search result
var searchResult = new searchResultObject();
searchResult.init(responseDetails);
//first check to see if you got locked out of google.
if(searchResult.pageTitle() == '403 Forbidden')
{
var lockOutNotice = ' <a href="' + searchURL + '">' + 'Lockout! - 1)Delete your google.com cookie, 2)click here and enter in the phrase, 3)should work now' + '</a>';
modifyPubmedResult(resultLine,lockOutNotice,'lockout');
}
else //carry on with the rest of the search
{
//what to use as the marker for the search being done
var doneComment = 'done';
if(searchMode == 'greek') doneComment = '^greek search done';
//if this the second search after inserting the greek characters, then remove the temporary notice saying that it's re-searching
if(searchMode == 'greek') modifyPubmedResult(resultLine,greekReSearchNotice,'','replace');
//if this the third search author search, then remove the temporary notice saying that it's re-searching
if(searchMode == 'author') modifyPubmedResult(resultLine,authorReSearchNotice,'','replace');
if (searchResult.hasResults())
{
//if there are multiple results, then compare the actual pubmed result to each google result
var bestMatchIndex = 0;
var bestMatch;
var multipleResults = false;
if(searchResult.getNumResults() > 1)
{
bestMatchIndex = searchResult.matchTitles(resultLine);
bestMatch = searchResult.getBestMatch();
multipleResults = true;
//modifyPubmedResult(resultLine,' <a href="' + searchURL + '">' + '(Multiple search results - ' + searchResult.getNumResults() + '), using ' + bestMatch + '% match hit</a>');
}
//if there were multiple results, bestMatchIndex is set to the index of the best google result match, otherwise the default is zero, or the first(and only) result
if(searchResult.getCitedByLink(bestMatchIndex))
{
modifyPubmedResult(resultLine,searchResult.getCitedByLink(bestMatchIndex),doneComment);
//debug if(searchMode == 'greek') modifyPubmedResult(resultLine,' (after greek fix)','^greek search done');
//debug if(searchMode == 'author') modifyPubmedResult(resultLine,' (after author fix)');
if(multipleResults) modifyPubmedResult(resultLine,' <a href="' + searchURL + '">' + '(' + bestMatch + '% hit)</a>');
//modifyPubmedResult(resultLine,' <a href="' + searchURL + '">' + '(Multiple search results - ' + searchResult.getNumResults() + '), using ' + bestMatch + '% match hit</a>');
}
else //if we found a result, but there are no citations
{
var noCitesLink = ' <a href="' + searchURL + '">' + 'No citations - see result' + '</a>';
modifyPubmedResult(resultLine,noCitesLink,doneComment);
//debug if(searchMode == 'greek') modifyPubmedResult(resultLine,' (after greek fix)','^greek search done');
//debug if(searchMode == 'author') modifyPubmedResult(resultLine,' (after author fix)');
if(multipleResults) modifyPubmedResult(resultLine,' <a href="' + searchURL + '">' + '(' + bestMatch + '% hit)</a>');
}
}
//if no search results were found, then one reason could be that Pubmed spelled out a greek letter in the title, but Google stores it as the greek letter itself
//so, check if there are any spelled out greek letters in the title, then tell the program to re-search them
else if((thisTitle.match('alpha')) || (thisTitle.match('beta')) || (thisTitle.match('gamma')) || (thisTitle.match('delta')) || (thisTitle.match('kappa')))
{
//add this resultLine to the list of ones to search for later
badTitleResultLines[badGreekTitles] = resultLine;
//increment the index of greek titles for the next one
badGreekTitles++;
modifyPubmedResult(resultLine,greekReSearchNotice,doneComment);
}
else if(searchMode == 'author') // if this was an author search, then it was the last effort, no results found //if nothing else, then link to a google search for the hell of it ----- new feature: next version have this part store the result line for doing an author search
{
var noResultsLink = ' <a href="' + searchURL + '">' + 'No results - see search' + '</a>';
modifyPubmedResult(resultLine,noResultsLink,doneComment);
//debug if(searchMode == 'author') modifyPubmedResult(resultLine,' (after author fix)');
}
else // for the results where title alone, or greek-fixed title didn't work, then save them to do an author search next
{
authorReSearchLines[authorReSearches] = resultLine;
authorReSearches++;
modifyPubmedResult(resultLine,authorReSearchNotice,doneComment);
//debug if(searchMode == 'greek') modifyPubmedResult(resultLine,' (after greek fix)','^greek search done');
}
}
}
});
}
function modifyPubmedResult(divIndex,linkText,commentText,type) //adds in the specified text to the search result; if 'type' is specifice as 'replace' then will remove the specified test
{
if(type == 'replace')
{
document.getElementsByTagName('div')[divIndex+2].innerHTML = document.getElementsByTagName('div')[divIndex+2].innerHTML.replace(linkText,'');
}
else
{
document.getElementsByTagName('div')[divIndex+2].innerHTML += linkText + '<!--' + commentText + '-->';
}
}
function searchResultObject() //an object containing info for each Google scholar search result
{
this.init = _init;
this.getNumResults = _getNumResults;
this.pageTitle = _pageTitle;
this.hasResults = _hasResults;
this.getCitedByLink = _getCitedByLink;
this.matchTitles = _matchTitles;
this.getBestMatch = _getBestMatch;
var searchDocument;
var numDivs;
var numResults;
var topMatch = 0;
var topMatchIndex = 0;
function _init(details)
{
//got this neat trick for parsing the HTML in responseDetails from http://blogs.pathf.com/agileajax/bjax/index.html????
searchDocument = document.createElement(searchDocument);
// do the XHR thing...
searchDocument.innerHTML = details.responseText; // contains the full html of a page
// voila, div now is the Root of an HTML DOM tree that can be traversed for screen scraping
numDivs = searchDocument.getElementsByTagName('div').length;
}
function _getNumResults()
{
if(searchDocument.getElementsByTagName('table')[3].getElementsByTagName('b')[4].previousSibling.textContent.substr(1,2) == 'of')
{
numResults = searchDocument.getElementsByTagName('table')[3].getElementsByTagName('b')[4].textContent;
}
return numResults;
}
function _pageTitle() //returns the title of the search result page
{
return searchDocument.getElementsByTagName('title')[0].textContent;
}
function _hasResults() //if the result page has a <div> element then there are results
{
if(numDivs >= 1)
{
return true;
}
else
{
return false;
}
}
function _getBestMatch()
{
return topMatch;
}
function _matchTitles(resultLine)
{
var firstResultIndex = 0;
if(searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[0].textContent.substr(0,12) == 'Did you mean')
{
firstResultIndex++;
}
var numResults = searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p').length - firstResultIndex;
var titleMatches = new Array();
var titleCounter = 0;
for (var i = firstResultIndex; i < numResults; i++)
{
var wordMatches = 0;
var thisResultTitle = searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[i].getElementsByTagName('a')[0].textContent.toLowerCase();
var actualTitle = allResults[resultLine].toLowerCase();
var thisResultTitleWords = thisResultTitle.split(' ');
var actualTitleWords = actualTitle.split('+');
//debug GM_log('thisResultTitle ' + thisResultTitle + 'actualTitle ' + actualTitle + 'thisResultTitleWords ' + thisResultTitleWords + ' actualTitleWords ' + actualTitleWords);
for(var at = 0; at < actualTitleWords.length; at++)
{
for(var trt = 0; trt < thisResultTitleWords.length; trt++)
{
if(actualTitleWords[at] == thisResultTitleWords[trt])
{
wordMatches++;
break;
}
}
}
titleMatches[titleCounter] = Math.round(wordMatches / thisResultTitleWords.length * 100);
titleCounter++;
}
//debug GM_log(titleMatches);
for (var i = 0; i < titleMatches.length; i++)
{
if(titleMatches[i] > topMatch)
{
topMatch = titleMatches[i];
topMatchIndex = i;
}
}
if(topMatch > 100)
topMatch = 100;
return topMatchIndex;
}
function _getCitedByLink(bestTitleIndex)
{
var newCitedLink;
var firstResultIndex = 0;
if(searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[0].textContent.substr(0,12) == 'Did you mean')
{
firstResultIndex++;
}
//GM_log('result index: ' + firstResultIndex);
firstResultIndex += bestTitleIndex;
var numLinks = searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[firstResultIndex].getElementsByTagName('a').length;
for (var i = 0; i < numLinks; i++) //will iterate through the <a> links on the page; one of them is the cites
{
var citedText = searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[firstResultIndex].getElementsByTagName('a')[i].textContent;
// now check if each <a> link is the "Cited by", then modify the current page
if(citedText.substr(0,5) == 'Cited')
{
// pull out the link to the citing papers
var citedLink = searchDocument.getElementsByTagName('div')[0].getElementsByTagName('p')[firstResultIndex].getElementsByTagName('a')[i].href;
// for some reason, 'http://www.ncbi.nlm.nih.gov/' gets added to the "href" in place of 'http://scholar.google.com/' which is weird, but easy to fix
newCitedLink = citedLink.replace('http://www.ncbi.nlm.nih.gov/','');
// make up the new "Cited by" link, then insert it at the end of each pubmed reference.
newCitedLink = ' <a href="http://scholar.google.com/' + newCitedLink + '">' + citedText + '</a>';
break;
}
}
return newCitedLink;
}
}
function checkForMultipleScripts()
{
var dupComments = 0;
for(var i in allResults)
{
var resultText = document.getElementsByTagName('div')[parseInt(i)+2].textContent;
if(resultText.indexOf("Cited by") != -1){
if(resultText.indexOf("Cited by",resultText.indexOf("Cited by") + 1) != -1){
dupComments++;
//debug GM_log(resultText + ' ' + dupComments);
}}
else if(resultText.indexOf("No results") != -1){
if(resultText.indexOf("No results",resultText.indexOf("No results") + 1) != -1){
dupComments++;
//debug GM_log(resultText + ' ' + dupComments);
}}
else if(resultText.indexOf("No citations") != -1){
if(resultText.indexOf("No citations",resultText.indexOf("No citations") + 1) != -1){
dupComments++;
//debug GM_log(resultText + ' ' + dupComments);
}}
else if(resultText.indexOf("author search") != -1){
if(resultText.indexOf("author search",resultText.indexOf("author search") + 1) != -1){
dupComments++;
//debug GM_log(resultText + ' ' + dupComments);
}}
else if(resultText.indexOf("much Greek") != -1){
if(resultText.indexOf("much Greek",resultText.indexOf("much Greek") + 1) != -1){
dupComments++;
//debug GM_log(resultText + ' ' + dupComments);
}}
}
//debug GM_log('dupcomments: ' + dupComments);
if(dupComments > Math.round(allResultsIndex/3)){
alert('It looks like you have multiple versions of this script running. To delete the old one, go to go to Tools -> Greasemonkey -> Manage User Scripts..., then select the old script and click the Uninstall button at the bottom.');}
}
function checkForUpdates() //the autoupdater
{
GM_xmlhttpRequest(
{
method: 'GET',
url: 'http://userscripts.org/scripts/show/13704',
headers: {
'User-agent': 'Mozilla/4.0 (compatible) Greasemonkey/0.3',
'Accept': 'application/atom+xml,application/xml,text/xml',},
onload: function(responseDetails)
{
//got this neat trick for parsing the HTML in responseDetails from http://blogs.pathf.com/agileajax/bjax/index.html
var div = document.createElement(div);
// do the XHR thing...
div.innerHTML = responseDetails.responseText; // contains the full html of a page
// voila, div now is the Root of an HTML DOM tree that can be traversed for screen scraping
//use this instead of the title as the version delimiter so users don't have to uninstall the old script
var newVersion = div.getElementsByTagName('div')[3].getElementsByTagName('p')[0].textContent.replace('Current Version: ','');
//debug GM_log('new version: _' + newVersion + '_');
if(newVersion != thisVersion)
{
var x=window.confirm("There is a new version of PubMed Citations (v" + newVersion + ") available at Userscripts.org \n You are currently using version " + thisVersion + " \n \n Click \"Ok\" to download it.");
if (x) {window.location = 'http://userscripts.org/scripts/source/13704.user.js'}
}
}
});
}
