citeclaw 2.0.0

package/vendor/translators-official/The Boston Globe.js

@@ -0,0 +1,285 @@
+{
+	"translatorID": "1f245496-4c1b-406a-8641-d286b3888231",
+	"label": "The Boston Globe",
+	"creator": "Adam Crymble, Frank Bennett, Sebastian Karcher",
+	"target": "^https?://(www|search|articles|archive)\\.boston\\.com/",
+	"minVersion": "2.1.9",
+	"maxVersion": "",
+	"priority": 100,
+	"inRepository": true,
+	"translatorType": 4,
+	"browserSupport": "gcsibv",
+	"lastUpdated": "2025-05-01 14:47:55"
+}
+
+/*
+ * Sample URLs
+ *
+ * [Original request -- uncommon page format, no embedded metadata of any kind]
+ * http://articles.boston.com/2011-05-03/news/29500032_1_bouncer-assault-local-restaurant
+ *
+ * [More common page formats, marginally reliable metadata in a comment block]
+ * http://www.boston.com/yourtown/news/charlestown/2011/04/meet_charlestowns_youth_of_the.html
+ * http://www.boston.com/business/articles/2011/05/05/oil_drops_below_100_per_barrel/
+ * http://www.boston.com/lifestyle/articles/2011/04/28/anticipation_grows_for_mfas_art_in_bloom_festival/
+
+ * Support for search results will require rewriting scrape(..) to use only regular expressions
+ */
+
+function detectWeb(doc, url) {
+	if (doc.location.hostname === 'archive.boston.com' && /(\/[0-9]{4}\/[0-9]{2}\/|[0-9]{4}-[0-9]{2}-[0-9]{2})/.test(url)) {
+		return "newspaperArticle";
+	}
+	return false;
+}
+
+//Boston Globe and Boston.com Translator. Original code by Adam Crymble
+// Rewritten by Frank Bennett, 2011
+
+function sniffComment (elem) {
+	if (!elem) {
+		return elem;
+	}
+	for (var i = 0, ilen = elem.childNodes.length; i < ilen; i += 1) {
+		if (elem.childNodes[i].nodeName === "#comment") {
+			return elem.childNodes[i].nodeValue;
+		}
+	}
+	return false;
+}
+
+function findMagicComment (doc) {
+	var hideMeElems = doc.getElementsByClassName("hideMe");
+	for (var i = 0, ilen = hideMeElems.length; i < ilen; i += 1) {
+		var elem = hideMeElems.item(i);
+		var sniff = sniffComment(elem);
+		if (sniff) {
+			return sniff;
+		}
+	}
+	var contentElem = doc.getElementById("content");
+	return sniffComment(contentElem);
+}
+
+function findAuthorString (doc, newItem) {
+	var authors = "";
+	var bylineElem = false;
+	var bylineElems = doc.getElementsByClassName("byline");
+	if (bylineElems.length) {
+		bylineElem = bylineElems.item(0);
+	}
+	if (!bylineElem) {
+		var bylineElem = doc.getElementById('byline');
+	}
+	if (bylineElem) {
+		authors = bylineElem.textContent;
+		authors = authors.replace(/\n/g, " ");
+		if (authors.match(/[Pp]osted\s+by\s+/)) {
+			newItem.itemType = "blogPost";
+		}
+		authors = authors.replace(/^\s*(?:[Bb]y|[Pp]osted\s+by)\s+(.*)/, "$1");
+	}
+	return authors;
+}
+
+function scrape (doc, url) {
+	// The site content is pretty chaotic, we do our best.
+
+	// There are two independent blocks set-and-save blocks
+	// below.
+
+	// Many pages seem to have metadata embedded in a comment
+	// The date and headline info look reliable, but
+	// the byline is a disaster, to be used only
+	// if absolutely necessary.
+	var magicComment = findMagicComment(doc);
+	if (magicComment) {
+		// Blind acceptance
+		var newItem =new Zotero.Item("newspaperArticle");
+		newItem.publicationTitle = "Boston.com";
+		// URL	
+		newItem.url = doc.location.href;
+		// Attachment
+		newItem.attachments.push({url:doc.location.href,mimeType:"text/html",snapshot:true,title:"Boston.com page"});
+		// Now try to get some citation details (go ahead, try)
+		var info = magicComment.replace('\n','','g');
+		newItem.title = ZU.xpathText(doc, '//div[@id="headTools"]/h1');
+		newItem.date = ZU.xpathText(doc, '//span[@id="dateline"]/text()[2]');
+		var authors = findAuthorString(doc, newItem);
+		if (!authors) {
+			var authors = info.replace(/.*<byline>(.*)<\/byline>.*/,"$1");
+			if (authors.toLowerCase() === authors) {
+				authors = info.replace(/.*<teasetext>(.*)<\/teasetext>.*/, "$1");
+				var m = authors.match(/^(?:[Bb]y\s+)*([^ ,]+).*/);
+				if (m) {
+					authors = m[1];
+				} else {
+					authors = "";
+				}
+			}
+		}
+		authors = authors.split(/,*\s+and\s+/);
+		authors[authors.length - 1] = authors[authors.length - 1].split(/,\s+/)[0];
+		authors = authors.join(", ");
+		authors = authors.split(/,\s+/);
+		for (var j = 0, jlen = authors.length; j < jlen; j += 1) {
+			var author = Zotero.Utilities.cleanAuthor(authors[j], 'author');
+			if (author.lastName) {
+				newItem.creators.push(author);
+			}
+		}
+		newItem.complete();
+	}
+
+
+	// Information block
+	var infoElem = doc.getElementById("mod-article-byline");
+	if (infoElem) {
+		var newItem = new Zotero.Item("newspaperArticle");
+		newItem.publicationTitle = "Boston.com";
+		// URL	
+		newItem.url = doc.location.href;
+		newItem.attachments.push({url:doc.location.href,mimeType:"text/html",snapshot:true,title:"Boston.com page"});
+
+		// Date
+		var dateElem = infoElem.getElementsByClassName('pubdate');
+		if (dateElem.length) {
+			newItem.date = dateElem.textContent;
+		}
+
+		// Authors
+		/*
+		for (var i = 0, ilen = infoElem.childNodes.length; i < ilen; i += 1) {
+			var node = infoElem.childNodes.item(i);
+			if (node.nodeName === 'SPAN') {
+				if ('By' === node.textContent.slice(0,2)) {
+					
+					var authors = node.textContent.slice(3);
+					authors = authors.split(/(?:, |,*\s+and\s+)/);
+					for (var j = 0, jlen = authors.length; j < jlen; j += 1) {
+						var author = Zotero.Utilities.cleanAuthor(authors[j], 'author');
+						newItem.creators.push(author);
+					}
+				}
+			}
+		}*/
+		
+		var authors = ZU.xpathText(infoElem, './span[@class="separator"]/following-sibling::span')
+		authors = authors.replace(/^\s*[Bb]y|,.+?$/g, "").trim();
+		author = authors.split(/ and |\s*,\s*/)
+		for (var i in author){
+			newItem.creators.push(ZU.cleanAuthor(author[i], "author"));
+		}
+		
+		// Title	
+		var headerElem = doc.getElementById('mod-article-header');
+		if (headerElem) {
+			var h = headerElem.getElementsByTagName('h1');
+			if (h.length) {
+				newItem.title = h[0].textContent;
+			}
+		}
+		newItem.complete();
+	}
+}
+
+
+function doWeb (doc, url) {
+
+	var articles = new Array();
+
+	if (detectWeb(doc, url) == "multiple") {
+		var items = {};
+		var result =  doc.evaluate('//div[@class="regTZ"]/a[@class="titleLink"]', doc, null, XPathResult.ANY_TYPE, null);
+		var elmt = result.iterateNext();
+		while (elmt) {
+			//items.push(elmt.href);
+			items[elmt.href] = elmt.textContent;
+			elmt = result.iterateNext();
+		}
+		Zotero.selectItems(items, function (items) {
+			if (!items) {
+				return true;
+			}
+			for (var i in items) {
+				articles.push(i);
+			}
+			ZU.processDocuments(articles, scrape);
+		});
+	} else {
+		scrape(doc, url);
+	}
+}
+
+/** BEGIN TEST CASES **/
+var testCases = [
+	{
+		"type": "web",
+		"url": "http://archive.boston.com/lifestyle/articles/2011/04/28/anticipation_grows_for_mfas_art_in_bloom_festival/?camp=pm",
+		"items": [
+			{
+				"itemType": "newspaperArticle",
+				"title": "Anticipation grows for MFA’s spring flower festival",
+				"creators": [
+					{
+						"firstName": "Carol",
+						"lastName": "Stocker",
+						"creatorType": "author"
+					}
+				],
+				"date": "April 28, 2011",
+				"libraryCatalog": "The Boston Globe",
+				"publicationTitle": "Boston.com",
+				"url": "http://archive.boston.com/lifestyle/articles/2011/04/28/anticipation_grows_for_mfas_art_in_bloom_festival/?camp=pm",
+				"attachments": [
+					{
+						"mimeType": "text/html",
+						"snapshot": true,
+						"title": "Boston.com page"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	},
+	{
+		"type": "web",
+		"url": "http://archive.boston.com/news/nation/washington/articles/2011/05/08/bin_laden_occupied_shrunken_dark_world/",
+		"items": [
+			{
+				"itemType": "newspaperArticle",
+				"title": "A peek inside bin Laden’s world: isolation, vanity, power",
+				"creators": [
+					{
+						"firstName": "Elisabeth",
+						"lastName": "Bumiller",
+						"creatorType": "author"
+					},
+					{
+						"firstName": "Carlotta",
+						"lastName": "Gall",
+						"creatorType": "author"
+					}
+				],
+				"date": "May 8, 2011",
+				"libraryCatalog": "The Boston Globe",
+				"publicationTitle": "Boston.com",
+				"shortTitle": "A peek inside bin Laden’s world",
+				"url": "http://archive.boston.com/news/nation/washington/articles/2011/05/08/bin_laden_occupied_shrunken_dark_world/",
+				"attachments": [
+					{
+						"mimeType": "text/html",
+						"snapshot": true,
+						"title": "Boston.com page"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	}
+]
+/** END TEST CASES **/

package/vendor/translators-official/The Chronicle of Higher Education.js

@@ -0,0 +1,308 @@
+{
+	"translatorID": "1e6d1529-246f-4429-84e2-1f1b180b250d",
+	"label": "The Chronicle of Higher Education",
+	"creator": "Simon Kornblith, Avram Lyon",
+	"target": "^https?://(www\\.)?chronicle\\.com/",
+	"minVersion": "2.1",
+	"maxVersion": "",
+	"priority": 100,
+	"inRepository": true,
+	"translatorType": 4,
+	"browserSupport": "gcsbv",
+	"lastUpdated": "2017-06-30 05:37:23"
+}
+
+/*
+ This translator works on articles posted in The Chronicle of Higher Education.
+
+ It is based on the earlier translator by Simon Kornblith, but the Chronicle has
+ significantly restructured the site since 2006, breaking the old translator.
+*/
+
+
+function detectWeb(doc, url) {
+	/* The /daily/ and /weekly/ sections are leftover from the previous version
+	   of the translator; they don't appear to still be on the Chronicle site, but
+	   they might persist in older URLs. */
+	var articleRegexp = /\/(daily|weekly|article|blogPost|blogs\/\w+)\/[^/]+\// ;
+	if (articleRegexp.test(url) && ZU.xpathText(doc, '//h1')) {
+		var section = url.match(articleRegexp);
+		switch (section[1]) {
+			case "weekly":
+			case "daily":
+			case "article":
+				return "magazineArticle";
+			case "blogPost":    
+				return "blogPost";
+			default:
+				if (section[1].indexOf("blogs") !== -1)
+					return "blogPost";
+				return false;
+		}
+	} else if (getSearchResults(doc, true)) {
+		return "multiple";
+	}
+}
+
+
+function getSearchResults(doc, checkOnly) {
+	var items = {};
+	var found = false;
+	//search results
+	var rows = ZU.xpath(doc, '//span[@class="content-card__heading"]/h4/a');
+	if (rows.length<1){
+		//overview pages
+		rows = ZU.xpath(doc, '//div[contains(@class, "row")]//h1/a|//div[contains(@class, "row")]//h2[contains(@class, "title")]/a');
+	}
+	for (var i=0; i<rows.length; i++) {
+		var href = rows[i].href;
+		var title = ZU.trimInternal(rows[i].textContent);
+		if (!href || !title) continue;
+		if (checkOnly) return true;
+		found = true;
+		items[href] = title;
+	}
+	return found ? items : false;
+}
+
+
+function doWeb(doc, url) {
+	if (detectWeb(doc, url) == "multiple") {
+		Zotero.selectItems(getSearchResults(doc, false), function (items) {
+			if (!items) {
+				return true;
+			}
+			var articles = [];
+			for (var i in items) {
+				articles.push(i);
+			}
+			ZU.processDocuments(articles, scrape);
+		});
+	} else {
+		scrape(doc, url);
+	}
+}
+
+
+function scrape (doc, url){
+		var type = detectWeb(doc, url);
+		var item = new Zotero.Item(type);
+
+		item.url = ZU.xpathText(doc, '//link[@rel="canonical"]/@href');
+		item.publicationTitle = "The Chronicle of Higher Education";
+		// Does the ISSN apply to online-only blog posts?
+		item.ISSN = "0009-5982";
+		item.language = "en-US";
+		
+		var byline = doc.evaluate('//header/div/span[@class="content-item__byline"]', doc, null, XPathResult.ANY_TYPE, null).iterateNext();
+		if (!byline) byline = doc.evaluate('//div[@class="blog__author"]/a', doc, null, XPathResult.ANY_TYPE, null).iterateNext();
+		if (byline !== null) {
+			var authors = parseAuthors(byline.textContent.trim());
+			for (var i = 0; i < authors.length; i++) {
+				item.creators.push(Zotero.Utilities.cleanAuthor(authors[i], "author"));
+			}
+		}
+		
+		// Behavior for some items is different:
+		if (type === "blogPost") {
+			item.date = ZU.xpathText(doc, '//div[@class="blog__author"]/time');
+			item.title = ZU.xpathText(doc, '//h2[@class="blog__title"]');
+			//legacy blogs
+			if (!item.title) item.title = ZU.xpathText(doc, '//h1[@class="title"]')
+			if (!item.date) item.date= ZU.xpathText(doc, '//p[@class="time"]');
+			
+			var blogname = ZU.xpathText(doc, '//div[@class="blog__mast"]//h2[contains(@class, "blog__name")]');
+			if (blogname) item.publicationTitle = item.publicationTitle + " Blogs: " + blogname;
+		} else {
+			item.date = ZU.xpathText(doc, '//header/div/span[@class="content-item__date"]')
+			item.title = ZU.xpathText(doc, '//header/h1');
+			var section = ZU.xpathText(doc, '//div[@class="header-breadcrumb-wrap"]/h1');
+			if (section) item.section = ZU.trimInternal(section)
+			
+			// Some items have publication details at the end of the article; one
+			// example is: http://chronicle.com/article/Grinnells-Green-Secrets/2653/
+			var articleParagraphs = doc.evaluate('//div[@class="article-body"]/p', doc, null, XPathResult.ANY_TYPE, null);
+			var par;
+			while ((par = articleParagraphs.iterateNext()) !== null) {
+				var data = par.textContent.match(/Section: ([a-zA-Z -&]+)[\n\t ]+Volume ([0-9]+), Issue ([0-9]+), Page ([0-9A-Za-z]+)/);
+				if (data !== null && data.length > 0) {
+					item.pages = data[4];
+					// If the section here and in the page proper are different, concatenate
+					if (item.section !== data[1])
+						item.section = item.section + " : " + Zotero.Utilities.trimInternal(data[1]);
+					// Since newspaperArticle doesn't have Volume / Issue, put as Edition
+					item.edition = "Volume " + data[2] + ", Issue " + data[3];
+				}
+			}
+		}
+		
+		item.attachments.push({url:doc.location.href, title: ("Chronicle of Higher Education Snapshot"), mimeType:"text/html"});
+		item.complete();
+
+}
+
+function parseAuthors(author) {
+		// Sometimes we have "By Author and Author"
+		if (author.substr(0, 3).toLowerCase() == "by ") {
+			author = author.substr(3);
+		}
+		
+		// Sometimes the author is in all caps
+		var pieces = author.split(" ");
+		for (var i = 0; i < pieces.length; i++) {
+			// TODO Make the all-caps character class more inclusive
+			if (pieces[i].match(/[A-Z-]+/) !== null)
+				pieces[i] = Zotero.Utilities.capitalizeTitle(pieces[i].toLowerCase(), true);
+		}
+		author = pieces.join(" ");
+		
+		// Somtimes we have multiple authors
+		var authors = author.split(" and ");
+		return authors;
+}
+
+
+/** BEGIN TEST CASES **/
+var testCases = [
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/blogs/profhacker/the-second-day-of-thatcamp/23068",
+		"items": [
+			{
+				"itemType": "blogPost",
+				"title": "The Second Day of THATCamp",
+				"creators": [
+					{
+						"firstName": "Amy",
+						"lastName": "Cavender",
+						"creatorType": "author"
+					}
+				],
+				"date": "March 26, 2010",
+				"blogTitle": "The Chronicle of Higher Education Blogs: ProfHacker",
+				"language": "en-US",
+				"url": "http://www.chronicle.com/blogs/profhacker/the-second-day-of-thatcamp/23068",
+				"attachments": [
+					{
+						"title": "Chronicle of Higher Education Snapshot",
+						"mimeType": "text/html"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	},
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/article/A-Little-Advice-From-32000/46210/",
+		"items": [
+			{
+				"itemType": "magazineArticle",
+				"title": "A Little Advice From 32,000 Graduate Students",
+				"creators": [
+					{
+						"firstName": "Adam",
+						"lastName": "Fagen",
+						"creatorType": "author"
+					},
+					{
+						"firstName": "Kimberly Suedkamp",
+						"lastName": "Wells",
+						"creatorType": "author"
+					}
+				],
+				"date": "January 14, 2002",
+				"ISSN": "0009-5982",
+				"language": "en-US",
+				"libraryCatalog": "The Chronicle of Higher Education",
+				"publicationTitle": "The Chronicle of Higher Education",
+				"url": "http://www.chronicle.com/article/A-Little-Advice-From-32000/46210",
+				"attachments": [
+					{
+						"title": "Chronicle of Higher Education Snapshot",
+						"mimeType": "text/html"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	},
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/article/Grinnells-Green-Secrets/2653/",
+		"items": [
+			{
+				"itemType": "magazineArticle",
+				"title": "Grinnell's Green Secrets",
+				"creators": [
+					{
+						"firstName": "Xiao-Bo",
+						"lastName": "Yuan",
+						"creatorType": "author"
+					}
+				],
+				"date": "June 16, 2006",
+				"ISSN": "0009-5982",
+				"language": "en-US",
+				"libraryCatalog": "The Chronicle of Higher Education",
+				"publicationTitle": "The Chronicle of Higher Education",
+				"url": "http://www.chronicle.com/article/Grinnells-Green-Secrets/2653",
+				"attachments": [
+					{
+						"title": "Chronicle of Higher Education Snapshot",
+						"mimeType": "text/html"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	},
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/blogs/brainstorm/humanities-cyberinfrastructure-project-bamboo/6138",
+		"items": [
+			{
+				"itemType": "blogPost",
+				"title": "Humanities Cyberinfrastructure: Project Bamboo",
+				"creators": [
+					{
+						"firstName": "Stan",
+						"lastName": "Katz",
+						"creatorType": "author"
+					}
+				],
+				"date": "July 17, 2008",
+				"blogTitle": "The Chronicle of Higher Education Blogs: Brainstorm",
+				"language": "en-US",
+				"shortTitle": "Humanities Cyberinfrastructure",
+				"url": "http://www.chronicle.com/blogs/brainstorm/humanities-cyberinfrastructure-project-bamboo/6138",
+				"attachments": [
+					{
+						"title": "Chronicle of Higher Education Snapshot",
+						"mimeType": "text/html"
+					}
+				],
+				"tags": [],
+				"notes": [],
+				"seeAlso": []
+			}
+		]
+	},
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/section/Opinion-Ideas/40/?eio=58977",
+		"items": "multiple"
+	},
+	{
+		"type": "web",
+		"url": "http://www.chronicle.com/search/?search_siteId=5&contextId=&action=rem&searchQueryString=adjunct",
+		"items": "multiple"
+	}
+]
+/** END TEST CASES **/