From f196cc5253eb02363739b23c11b733b83fbdb292 Mon Sep 17 00:00:00 2001 From: Michael Berkowitz Date: Mon, 14 Jul 2008 18:21:01 +0000 Subject: [PATCH] -Adds Revues.org translator. --- scrapers.sql | 58 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 58 insertions(+) diff --git a/scrapers.sql b/scrapers.sql index df34b322f..ac20daa68 100644 --- a/scrapers.sql +++ b/scrapers.sql @@ -1264,6 +1264,64 @@ function doWeb(doc, url) { Zotero.wait(); }'); +REPLACE INTO translators VALUES ('87766765-919e-4d3b-9071-3dd7efe984c8', '1.0.0b4.r5', '', '2008-07-14 14:19:40', '0', '100', '4', 'Revues.org', 'Michael Berkowitz', 'http://.*\.revues\.org', +'function detectWeb(doc, url) { + if (doc.evaluate(''//div[@id="inside"]/div[@class="sommaire"]/dl[@class="documents"]/dd[@class="titre"]/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext() + || doc.evaluate(''//ul[@class="summary"]//div[@class="title"]/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) { + return "multiple"; + } else if (doc.evaluate(''//h1[@id="docTitle"]/span[@class="text"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) { + return "journalArticle"; + } +}', +'function doWeb(doc, url) { + var arts = new Array(); + if (detectWeb(doc, url) == "multiple") { + if (doc.evaluate(''//ul[@class="summary"]//div[@class="title"]/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) { + var xpath = ''//ul[@class="summary"]//div[@class="title"]/a''; + } else if (doc.evaluate(''//div[@id="inside"]/div[@class="sommaire"]/dl[@class="documents"]/dd[@class="titre"]/a'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) { + var xpath = ''//div[@id="inside"]/div[@class="sommaire"]/dl[@class="documents"]/dd[@class="titre"]/a''; + } + var titles = doc.evaluate(xpath, doc, null, XPathResult.ANY_TYPE, null); + var title; + var items = new Object(); + while (title = titles.iterateNext()) { + items[title.href] = title.textContent; + } + items = Zotero.selectItems(items); + for (var i in items) { + arts.push(i); + } + } else { + arts = [url]; + } + Zotero.Utilities.processDocuments(arts, function(doc) { + var metas = doc.evaluate(''//meta'', doc, null, XPathResult.ANY_TYPE, null); + var meta; + var data = new Object(); + while (meta = metas.iterateNext()) { + if (data[meta.name]) { + data[meta.name] = data[meta.name] + ";" + meta.content; + } else { + data[meta.name] = meta.content + } + } + var item = new Zotero.Item("journalArticle"); + item.url = data[''url'']; + var authors = data[''author''].split('';''); + for each (var aut in authors) { + if (aut.match(/\w+/)) item.creators.push(Zotero.Utilities.cleanAuthor(aut.replace(/(.*)\s([^\s]+)$/, "$2 $1"), "author")); + } + item.tags = data[''DC.subject''].split(/,\s+/); + item.date = data[''DC.date'']; + item.title = data[''DC.title'']; + item.publicationTitle = data[''DC.relation.isPartOf''].match(/^[^,]+/)[0]; + item.abstractNote = data[''DC.description'']; + + item.complete(); + }, function() {Zotero.done;}); + Zotero.wait(); +}'); + REPLACE INTO translators VALUES ('daa26181-71d4-48ef-8cac-54c06ff4c20e', '1.0.0b4.r5', '', '2008-07-07 23:40:00', '0', '100', '4', 'Citebase', 'Michael Berkowitz', 'http://(www.)?citebase.org/', 'function detectWeb(doc, url) { if (url.match(/\/search/)) {