Ticket #918: ACM.txt

var urls = new Array();

// this handles sequential loading, since first we need to process a document (to get the abstract), then
// get the Refer metadata, then process the next document, etc.
function getNext() {
        if(urls.length) {
                var url = urls.shift();
                Zotero.Utilities.processDocuments([url], function(doc) { scrape(doc); });
        } else {
                Zotero.done();
        }
}

function scrape(doc) {
        var onClick = doc.evaluate('//a[substring(text(), 5, 7) = "EndNote"]', doc, null, XPathResult.ANY_TYPE,
                null).iterateNext().getAttribute("onClick");
        var m = onClick.match(/'([^']+)'/);
        
        var abstract = doc.evaluate('//div[@class="abstract"]/p[@class="abstract"]', doc, null,
                XPathResult.ANY_TYPE, null).iterateNext();
        if(abstract) abstract = Zotero.Utilities.cleanString(abstract.textContent);
        
        var snapshot = doc.location.href;
        
        var attachments = new Array();
        var url;
        var typeLinks = doc.evaluate('//td[@class="smaller-text"]/a[img]', doc, null,
                XPathResult.ANY_TYPE, null);
        var typeLink;
        while(typeLink = typeLinks.iterateNext()) {
                var linkText = typeLink.textContent.toLowerCase();
                if(linkText == "pdf") {
                        attachments.push({title:"ACM Full Text PDF", mimeType:"application/pdf", url:typeLink.href});
                        url = typeLink.href;
                } else if(linkText == "html") {
                        url = snapshot = typeLink.href;
                }
        }
        
        attachments.push({title:"ACM Snapshot", mimeType:"text/html", url:snapshot});

        var doi;
        var doiIter;
        var doiList = doc.evaluate('//td[@class="small-text"]/a', doc, null, XPathResult.ANY_TYPE, null);
        while (doiIter = doiList.iterateNext()) {
        var linkText = doiIter.textContent;
        if (linkText.match(/doi\.acm\.org/)) {
                doi = linkText.match(/doi\.acm\.org\/(.*)/)[1];
                }
        }

        var keywords = new Array();
        var keywordLinks = doc.evaluate('//p[@class="keywords"]/a', doc, null,
                XPathResult.ANY_TYPE, null);
        var keywordLink;
        while(keywordLink = keywordLinks.iterateNext()) {
                keywords.push(Zotero.Utilities.trim(keywordLink.textContent.toLowerCase()));
        }
        
        Zotero.Utilities.HTTP.doGet("http://portal.acm.org/"+m[1], function(text) {
                // split() may no longer be necessary
                var m = text.split(/<\/?pre[^>]*>/ig);
                if (m[1]) {
                        var text = m[1];
                }
                
                // load Refer translator
                var translator = Zotero.loadTranslator("import");
                translator.setTranslator("881f60f2-0802-411a-9228-ce5f47b64c7d");
                translator.setString(text);
                translator.setHandler("itemDone", function(obj, item) {
                        if(abstract) item.abstractNote = abstract;
                        item.attachments = attachments;
                        item.tags = keywords;
                        item.type = undefined;
                        item.DOI = doi;
                        item.complete();
                });
                translator.translate();
                
                getNext();
        });
}

function doWeb(doc, url) {
        if(url.indexOf("/results.cfm") != -1) {
                var items = Zotero.Utilities.getItemArray(doc, doc, '^https?://[^/]+/citation.cfm\\?[^#]+$');
                
                items = Zotero.selectItems(items);
                if(!items) return true;
                
                for(var url in items) {
                        urls.push(url);
                }
                
                getNext();
        } else {
                scrape(doc);
        }
        
        Zotero.wait();
}