From 0ebf4ebedef54ecb04b9ed61abe5f210e92de27b Mon Sep 17 00:00:00 2001 From: Michael Berkowitz Date: Mon, 21 Apr 2008 17:46:08 +0000 Subject: [PATCH] -Adds Progress of Theoretical Physics translator. --- scrapers.sql | 69 +++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 68 insertions(+), 1 deletion(-) diff --git a/scrapers.sql b/scrapers.sql index aaec54431..aa937c2d7 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', '2008-04-21 15:30:00')); +REPLACE INTO version VALUES ('repository', STRFTIME('%s', '2008-04-21 17:45:00')); REPLACE INTO translators VALUES ('96b9f483-c44d-5784-cdad-ce21b984fe01', '1.0.0b4.r1', '', '2008-03-21 20:00:00', '1', '100', '4', 'Amazon.com', 'Sean Takats and Michael Berkowitz', '^https?://(?:www\.)?amazon', 'function detectWeb(doc, url) { @@ -2991,6 +2991,73 @@ REPLACE INTO translators VALUES ('62c0e36a-ee2f-4aa0-b111-5e2cbd7bb5ba', '1.0.0b } }'); +REPLACE INTO translators VALUES ('0b356cb6-7fa1-4662-b6e8-7ffc9ca2cd4a', '1.0.0b4.r5', '', '2008-04-21 17:45:00', '0', '100', '4', 'Progress of Theoretical Physics', 'Michael Berkowitz', 'http://ptp.ipap.jp/', +'function detectWeb(doc, url) { + if (doc.title.match(/search result/) || doc.title.match(/Table of Contents/)) { + return "multiple"; + } else if (url.match(/getarticle\?/)) { + return "journalArticle"; + } +}', +'function doWeb(doc, url) { + var arts = new Array(); + if (detectWeb(doc, url) == "multiple") { + var items = new Object(); + if (doc.title.toLowerCase().match(/search result/)) { + var titles = doc.evaluate(''/html/body//li//b'', doc, null, XPathResult.ANY_TYPE, null); + var links = doc.evaluate(''/html/body//li/a'', doc, null, XPathResult.ANY_TYPE, null); + var title; + var link; + while ((title = titles.iterateNext()) && (link = links.iterateNext())) { + items[link.href] = Zotero.Utilities.trimInternal(title.textContent); + } + } else if (doc.title.toLowerCase().match(/table of contents/)) { + var xpath = doc.evaluate(''/html/body/ul/li/a'', doc, null, XPathResult.ANY_TYPE, null); + var text; + while (text = xpath.iterateNext()) { + items[text.href] = Zotero.Utilities.trimInternal(text.textContent); + } + } + items = Zotero.selectItems(items); + for (var i in items) { + arts.push(i); + } + } else { + arts = [url]; + } + Zotero.Utilities.processDocuments(arts, function(doc) { + var item = new Zotero.Item("journalArticle"); + item.ISSN = ''0033-068X''; + item.title = Zotero.Utilities.trimInternal(doc.evaluate(''//h2[@class="title"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().textContent); + if (doc.evaluate(''//h2[@class="subtitle"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext()) { + item.title = item.title + ": " + Zotero.Utilities.trimInternal(doc.evaluate(''//h2[@class="subtitle"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().textContent); + } + item.abstractNote = Zotero.Utilities.trimInternal(doc.evaluate(''//p[@class="abstract"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().textContent); + var authors = Zotero.Utilities.unescapeHTML(Zotero.Utilities.trimInternal(doc.evaluate(''/html/body/p[@class="author"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().textContent)); + authors = authors.replace(/[*()]+/g, "").split(/(,\s+|\band\b)/); + for each (var aut in authors) { + if (!aut.match(/(,|and)/)) { + item.creators.push(Zotero.Utilities.cleanAuthor(aut, "author")); + } + } + var info = Zotero.Utilities.trimInternal(doc.evaluate(''//h4[@class="info"]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().textContent); + info = info.match(/Vol\.\s+(\d+)\s+No\.\s+(\d+)\s+\((\d+)\)\s+pp\.\s+([\d-]+)\s+URL\s+:\s+(.*)\s+DOI\s+:\s+(.*)$/); + item.volume = info[1]; + item.issue = info[2]; + item.date = info[3]; + item.pages = info[4]; + item.url = info[5]; + item.DOI = info[6]; + var pdfurl = doc.evaluate(''//a[contains(text(), "PDF")]'', doc, null, XPathResult.ANY_TYPE, null).iterateNext().href; + item.attachments = [ + {url:item.url, title:"PTP Snapshot", mimeType:"text/html"}, + {url:pdfurl, title:"PTP Full Text PDF", mimeType:"application/pdf"} + ]; + item.complete(); + }, function() {Zotero.done;}); + Zotero.wait(); +}'); + REPLACE INTO translators VALUES ('0863b8ec-e717-4b6d-9e35-0b2db2ac6b0f', '1.0.0b4.r5', '', '2008-04-21 15:30:00', '0', '100', '4', 'Institute of Pure and Applied Physics', 'Michael Berkowitz', 'http://(jjap|apex|jpsj)\.ipap\.jp/', 'function detectWeb(doc, url) { if (doc.title.indexOf("Table of Contents") != -1 || doc.title.indexOf("search result") != -1) {