From d906a4b9991ae52f7a388d94be9e2f3f77642e1d Mon Sep 17 00:00:00 2001 From: Michael Berkowitz Date: Fri, 14 Mar 2008 18:15:34 +0000 Subject: [PATCH] -Adds IOVS translator. --- scrapers.sql | 74 +++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 73 insertions(+), 1 deletion(-) diff --git a/scrapers.sql b/scrapers.sql index 56f29beca..7689f2888 100644 --- a/scrapers.sql +++ b/scrapers.sql @@ -22,7 +22,7 @@ -- Set the following timestamp to the most recent scraper update date -REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-03-14 18:00:00')); +REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-03-14 19:00:00')); REPLACE INTO translators VALUES ('96b9f483-c44d-5784-cdad-ce21b984fe01', '1.0.0b4.r1', '', '2007-06-21 20:00:00', '1', '100', '4', 'Amazon.com', 'Sean Takats', '^https?://(?:www\.)?amazon', 'function detectWeb(doc, url) { @@ -1579,6 +1579,78 @@ function doWeb(doc, url) { } }'); +REPLACE INTO translators VALUES ('4654c76f-451c-4ae6-9a36-575e982b3cdb', '1.0.0b4.r5', '', '2008-03-14 19:00:00', '0', '100', '4', 'Investigative Ophthalmology and Visual Science', 'Michael Berkowitz', 'http://www.iovs.org/', +'function detectWeb(doc, url) { + if (doc.title.indexOf("Table of Contents") != -1 || doc.title.indexOf("Search Result") != -1) { + return "multiple" + } else if (url.indexOf("abstract") != -1 || url.indexOf("full") != -1) { + return "journalArticle"; + } +}', +'function doWeb(doc, url) { + var host = doc.location.host; + var arts = new Array(); + if (detectWeb(doc, url) == "multiple") { + var items = new Object(); + if (doc.title.indexOf("Search Result") != -1) { + var boxes = doc.evaluate(''//table/tbody/tr/td/font/table/tbody/tr[1]'', doc, null, XPathResult.ANY_TYPE, null); + var box; + while (box = boxes.iterateNext()) { + var id = doc.evaluate(''.//input'', box, null, XPathResult.ANY_TYPE, null).iterateNext().value; + var titles = doc.evaluate(''./td/font/strong'', box, null, XPathResult.ANY_TYPE, null); + var titletext = ''''; + var title; + while (title = titles.iterateNext()) { + titletext += title.textContent; + } + items[id] = titletext; + } + } else if (doc.title.indexOf("Table of Content") != -1) { + var ids = doc.evaluate(''/html/body/form/dl/dt/input'', doc, null, XPathResult.ANY_TYPE, null); + var titles = doc.evaluate(''/html/body/form/dl/dd/strong'', doc, null, XPathResult.ANY_TYPE, null); + var id; + var title; + while ((title = titles.iterateNext()) && (id = ids.iterateNext())) { + items[''iovs;'' + id.value] = title.textContent; + } + } + items = Zotero.selectItems(items); + for (var i in items) { + arts.push(i); + } + } else { + arts = [doc.evaluate(''//a[contains(@href, "citmgr")]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().href.match(/=(.*)$/)[1]] + } + Zotero.debug(arts); + for each (var id in arts) { + var post = ''type=refman&gca='' + id; + Zotero.debug(post); + post = ''http://www.iovs.org/cgi/citmgr?'' + post; + Zotero.debug(post); + Zotero.Utilities.HTTP.doGet(post, function(text) { + Zotero.debug(text); + var translator = Zotero.loadTranslator("import"); + translator.setTranslator("32d59d2d-b65a-4da4-b0a3-bdd3cfb979e7"); + translator.setString(text); + translator.setHandler("itemDone", function(obj, item) { + var pdfurl = item.url.replace(/content\/[^/]+/, "reprint") + ".pdf"; + item.attachments = [ + {url:item.url, title:"IOVS Snapshot", mimeType:"text/html"}, + {url:pdfurl, tite:"IOVS Full Text PDF", mimeType:"application/pdf"} + ]; + if (item.notes[0][''note''].match(/\d/)) { + item.DOI = item.notes[0][''note'']; + item.notes = new Array(); + } + item.complete(); + }); + translator.translate(); + + Zotero.done(); + }); + } +}'); + REPLACE INTO translators VALUES ('62c0e36a-ee2f-4aa0-b111-5e2cbd7bb5ba', '1.0.0b4.r5', '', '2008-03-13 22:30:00', '0', '100', '4', 'MetaPress', 'Michael Berkowitz', 'https?://(.*).metapress.com/', 'function detectWeb(doc, url) { if (doc.title.indexOf("Search Results") != -1) {