CNKI.js

{
	"translatorID": "5c95b67b-41c5-4f55-b71a-48d5d7183063",
	"label": "CNKI",
	"creator": "lin xingzhong",
	"target": "^https?://([^/]+\\.)?cnki\\.net",
	"minVersion": "3.0",
	"maxVersion": "",
	"priority": 100,
	"inRepository": true,
	"translatorType": 4,
	"browserSupport": "gcs",
	"lastUpdated": "2019-11-18 01:34:23"
}

/*
	***** BEGIN LICENSE BLOCK *****

	CNKI(China National Knowledge Infrastructure) Translator
	Copyright © 2013 Aurimas Vinckevicius

	This file is part of Zotero.

	Zotero is free software: you can redistribute it and/or modify
	it under the terms of the GNU Affero General Public License as published by
	the Free Software Foundation, either version 3 of the License, or
	(at your option) any later version.

	Zotero is distributed in the hope that it will be useful,
	but WITHOUT ANY WARRANTY; without even the implied warranty of
	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	GNU Affero General Public License for more details.

	You should have received a copy of the GNU Affero General Public License
	along with Zotero. If not, see <http://www.gnu.org/licenses/>.

	***** END LICENSE BLOCK *****
*/

// fetches Refworks record for provided IDs and calls next with resulting text
// ids should be in the form [{dbname: "CDFDLAST2013", filename: "1013102302.nh"}]
function getRefworksByID(ids, next) {
	var postData = "";
	for (var i=0, n=ids.length; i<n; i++) {
		postData += ids[i].dbname + "!" + ids[i].filename + "!1!0,";
	}
	postData = "formfilenames=" + encodeURIComponent(postData);
	postData += '&hid_kLogin_headerUrl=/KLogin/Request/GetKHeader.ashx%3Fcallback%3D%3F';
	postData += '&hid_KLogin_FooterUrl=/KLogin/Request/GetKHeader.ashx%3Fcallback%3D%3F';
	postData += '&CookieName=FileNameS';
	
	ZU.doGet('https://kns.cnki.net/kns/ViewPage/viewsave.aspx?displayMode=Refworks' + '&' +  postData, 
		function(text) {
			var parser = new DOMParser();
			var html = parser.parseFromString(text, "text/html")
			var text = ZU.xpath(html, "//table[@class='mainTable']//td")[0].innerHTML;
			var text = text.replace(/<br>/g, '\n');
			text = text.replace(/^RT\s+Dissertation\/Thesis/gmi, 'RT Dissertation');
			text = text.replace(/^(A[1-4]|U2)\s*([^\r\n]+)/gm, 
					function(m, tag, authors) {
						var authors = authors.split(/\s*[;，,]\s*/); //that's a special comma
						if (!authors[authors.length-1].trim()) authors.pop();
						return tag + ' ' + authors.join('\n' + tag + ' ');
					}
				);
			next(text);
		}
	);
}

function getIDFromURL(url) {
	if (!url) return false;
	
	var dbname = url.match(/[?&]dbname=([^&#]*)/i);
	var filename = url.match(/[?&]filename=([^&#]*)/i);
	if (!dbname || !dbname[1] || !filename || !filename[1]) return false;
	
	return { dbname: dbname[1], filename: filename[1], url: url };
}


// 网络首发期刊信息并不能从URL获取dbname和filename信息
// Get dbname and filename from pre-released article web page.
function getIDFromRef(doc, url) {
	var func = ZU.xpath(doc, '//div[@class="link"]/a');
	if (!func.length) {
		return false;
	} else {
		func = func[0].onclick + '';
		var tmp = func.split(',')[1].split('!');
		// Z.debug(func + tmp[0].slice(1));
		return { dbname: tmp[0].slice(1), filename: tmp[1], url: url };
	}
}

function getIDFromPage(doc, url) {
	return getIDFromURL(url)
		|| getIDFromURL(ZU.xpathText(doc, '//div[@class="zwjdown"]/a/@href'))
		|| getIDFromRef(doc, url);
}

function getTypeFromDBName(dbname) {
	var dbType = {
		CJFQ: "journalArticle",
		CJFD: "journalArticle",
		CAPJ: "journalArticle",
		CDFD: "thesis",
		CMFD: "thesis",
		CLKM: "thesis",
		CCND: "newspaperArticle",
		CPFD: "conferencePaper",
	};
	var db = dbname.substr(0, 4).toUpperCase();
	if (dbType[db]) {
		return dbType[db];
	} else {
		return false;
	}
}

function getItemsFromSearchResults(doc, url, itemInfo) {
	var iframe = doc.getElementById('iframeResult');
	if (iframe) {
		var innerDoc = iframe.contentDocument || iframe.contentWindow.document;
		if (innerDoc) {
			doc = innerDoc;
		}
	}
	
	var links = ZU.xpath(doc, '//tr[not(.//tr) and .//a[@class="fz14"]]');
	var aXpath = './/a[@class="fz14"]';
	if (!links.length) {
		links = ZU.xpath(doc, '//table[@class="GridTableContent"]/tbody/tr[./td[2]/a]');
		aXpath = './td[2]/a';
	}

	if (!links.length) {
		return false;
	} else {
		var items = {};
		for (var i = 0, n = links.length; i < n; i++) {
			// Z.debug(links[i].innerHTML)
			var a = ZU.xpath(links[i], aXpath)[0];
			var title = ZU.xpathText(a, './node()[not(name()="SCRIPT")]', null, '');
			if (title) title = ZU.trimInternal(title);
			var id = getIDFromURL(a.href);
			// pre-released item can not get ID from URL, try to get ID from element.value
			if (!id) {
				var td1 = ZU.xpath(links[i], './td')[0];
				var tmp = td1.value.split('!');
				id = { dbname: tmp[0], filename: tmp[1], url: a.href };
			}
			Z.debug('links' + i);
			if (!title || !id) continue;

			if (itemInfo) {
				itemInfo[a.href] = { id: id };
			}
			items[a.href] = title;
		}

		return items;
	}
}

function detectWeb(doc, url) {
	// Z.debug(doc);
	var ab = ZU.xpath(doc, "//span[@id='ChDivSummary']")[0]
	var id = getIDFromPage(doc, url);
	var items = getItemsFromSearchResults(doc, url);
	Z.debug(id);
	if (id) {
		return getTypeFromDBName(id.dbname);
	} else if (items) {
		return "multiple";
	} else {
		return false;
	}
}

function doWeb(doc, url) {
	if (detectWeb(doc, url) == "multiple") {
		var itemInfo = {};
		var items = getItemsFromSearchResults(doc, url, itemInfo);
		Z.selectItems(items, function(selectedItems) {
			if (!selectedItems) return true;
			
			var itemInfoByTitle = {};
			var ids = [];
			for (var url in selectedItems) {
				ids.push(itemInfo[url].id);
				itemInfoByTitle[selectedItems[url]] = itemInfo[url];
				itemInfoByTitle[selectedItems[url]].url = url;
			}
			scrape(ids, doc, url, itemInfoByTitle);
		});
	} else {
		scrape([getIDFromPage(doc, url)], doc, url);
	}
}

function scrape(ids, doc, url, itemInfo) {
	getRefworksByID(ids, function(text, ids) {
		var resultNum = 'single';
		if (ids && ids.length > 1) {
			resultNum = 'multiple';
		}
		var translator = Z.loadTranslator('import');
		translator.setTranslator('1a3506da-a303-4b0a-a1cd-f216e6138d86'); //Refworks
		text = text.replace(/IS (\d+)\nvo/, "IS $1\nVO");
		translator.setString(text);
		
		translator.setHandler('itemDone', function(obj, newItem) {
			// split names
			for (var i = 0, n = newItem.creators.length; i < n; i++) {
				var creator = newItem.creators[i];
				if (creator.firstName) continue;
				
				var lastSpace = creator.lastName.lastIndexOf(' ');
				if (creator.lastName.search(/[A-Za-z]/) !== -1 && lastSpace !== -1) {
					// western name. split on last space
					creator.firstName = creator.lastName.substr(0,lastSpace);
					creator.lastName = creator.lastName.substr(lastSpace + 1);
				} else {
					// Chinese name. first character is last name, the rest are first name
					creator.firstName = creator.lastName.substr(1);
					creator.lastName = creator.lastName.charAt(0);
				}
			}
			
			if (newItem.abstractNote) {
				newItem.abstractNote = newItem.abstractNote.replace(/\s*[\r\n]\s*/g, '\n');
			}
			
			// clean up tags. Remove numbers from end
			for (var j = 0, l = newItem.tags.length; j < l; j++) {
				newItem.tags[j] = newItem.tags[j].replace(/:\d+$/, '');
			}
			
			newItem.title = ZU.trimInternal(newItem.title);
			if (itemInfo) {
				var info = itemInfo[newItem.title];
				if (!info) {
					Z.debug('No item info for "' + newItem.title + '"');
				} else {
					newItem.url = info.url;
				}
			} else {
				newItem.url = url;
			}

			// CN 中国刊物编号，非refworks中的callNumber
			// CN in CNKI refworks format explains Chinese version of ISSN
			if (newItem.callNumber){
			//	newItem.extra = 'CN ' + newItem.callNumber;
				newItem.callNumber = "";
			}
			// don't download PDF/CAJ on searchResult(multiple)
			if (resultNum == 'single'){
				newItem.attachments = getAttachments(doc, newItem);
			} else if (resultNum == 'multiple') {
				newItem.attachments = [{
					url: newItem.url,
					title: newItem.title,
					mimeType: "text/html",
					snapshot: true
				}];
			}
			newItem.complete();
		});
		
		translator.translate();
	});
}

// get pdf download link
function getPDF(doc) {
	var pdf = ZU.xpath(doc, "//a[@name='pdfDown']");
	return pdf.length ? pdf[0].href : false;
}

// caj download link, default is the whole article for thesis.
function getCAJ(doc, itemType) {
	// //div[@id='DownLoadParts']
	if (itemType == 'thesis') {
		var caj = ZU.xpath(doc, "//div[@id='DownLoadParts']/a");
	} else {
		caj = ZU.xpath(doc, "//a[@name='cajDown']");
	}
	return caj.length ? caj[0].href : false;
}

// add pdf or caj to attachments, default is pdf
function getAttachments(doc, item) {
	var attachments = [{
		url: item.url,
		title: item.title,
		mimeType: "text/html",
		snapshot: true
	}];
	var pdfurl = getPDF(doc);
	var cajurl = getCAJ(doc, item.itemType);
	// Z.debug('pdf' + pdfurl);
	// Z.debug('caj' + cajurl);
	var loginUser = ZU.xpath(doc, "//input[@id='loginuserid']");
	// Z.debug(doc.body.innerHTML);
	// Z.debug(loginUser[0].value);
	// Z.debug(loginUser.length);
	if (loginUser.length && loginUser[0].value) { 
		if (pdfurl) {
			attachments.push({
				title: "Full Text PDF",
				mimeType: "application/pdf",
				url: pdfurl
			});
		} else if (cajurl) {
			attachments.push({
				title: "Full Text CAJ",
				mimeType: "application/caj",
				url: cajurl
			});
		}
	}
	
	return attachments;
}

/** BEGIN TEST CASES **/
var testCases = [
	{
		"type": "web",
		"url": "http://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CJFQ&dbname=CJFDLAST2015&filename=SPZZ201412003&v=MTU2MzMzcVRyV00xRnJDVVJMS2ZidVptRmkva1ZiL09OajNSZExHNEg5WE5yWTlGWjRSOGVYMUx1eFlTN0RoMVQ=",
		"items": [
			{
				"itemType": "journalArticle",
				"title": "基于部分酸水解-亲水作用色谱-质谱的黄芪多糖结构表征",
				"creators": [
					{
						"lastName": "梁",
						"firstName": "图",
						"creatorType": "author"
					},
					{
						"lastName": "傅",
						"firstName": "青",
						"creatorType": "author"
					},
					{
						"lastName": "辛",
						"firstName": "华夏",
						"creatorType": "author"
					},
					{
						"lastName": "李",
						"firstName": "芳冰",
						"creatorType": "author"
					},
					{
						"lastName": "金",
						"firstName": "郁",
						"creatorType": "author"
					},
					{
						"lastName": "梁",
						"firstName": "鑫淼",
						"creatorType": "author"
					}
				],
				"date": "2014",
				"ISSN": "1000-8713",
				"abstractNote": "来自中药的水溶性多糖具有广谱治疗和低毒性特点,是天然药物及保健品研发中的重要组成部分。针对中药多糖结构复杂、难以表征的问题,本文以中药黄芪中的多糖为研究对象,采用\"自下而上\"法完成对黄芪多糖的表征。首先使用部分酸水解方法水解黄芪多糖,分别考察了水解时间、酸浓度和温度的影响。在适宜条件(4 h、1.5mol/L三氟乙酸、80℃)下,黄芪多糖被水解为特征性的寡糖片段。接下来,采用亲水作用色谱与质谱联用对黄芪多糖部分酸水解产物进行分离和结构表征。结果表明,提取得到的黄芪多糖主要为1→4连接线性葡聚糖,水解得到聚合度4~11的葡寡糖。本研究对其他中药多糖的表征具有一定的示范作用。",
				"issue": "12",
				"language": "中文;",
				"libraryCatalog": "CNKI",
				"pages": "1306-1312",
				"publicationTitle": "色谱",
				"url": "http://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CJFQ&dbname=CJFDLAST2015&filename=SPZZ201412003&v=MTU2MzMzcVRyV00xRnJDVVJMS2ZidVptRmkva1ZiL09OajNSZExHNEg5WE5yWTlGWjRSOGVYMUx1eFlTN0RoMVQ=",
				"volume": "32",
				"attachments": [
					{
						"title": "基于部分酸水解-亲水作用色谱-质谱的黄芪多糖结构表征",
						"mimeType": "text/html",
						"snapshot": true
					}
				],
				"tags": [
					{
						"tag": "Astragalus"
					},
					{
						"tag": "characterization"
					},
					{
						"tag": "hydrophilic interaction liquid chromatography(HILIC)mass spectrometry(MS)"
					},
					{
						"tag": "partial acid hydrolysis"
					},
					{
						"tag": "polysaccharides"
					},
					{
						"tag": "亲水作用色谱"
					},
					{
						"tag": "多糖"
					},
					{
						"tag": "表征"
					},
					{
						"tag": "质谱"
					},
					{
						"tag": "部分酸水解"
					},
					{
						"tag": "黄芪"
					}
				],
				"notes": [],
				"seeAlso": []
			}
		]
	},
	{
		"type": "web",
		"url": "https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201701&filename=1017045605.nh&v=MDc3ODZPZVorVnZGQ3ZrV3JyT1ZGMjZHYk84RzlmTXFwRWJQSVI4ZVgxTHV4WVM3RGgxVDNxVHJXTTFGckNVUkw=",
		"items": [
			{
				"itemType": "thesis",
				"title": "黄瓜共表达基因模块的识别及其特点分析",
				"creators": [
					{
						"lastName": "林",
						"firstName": "行众",
						"creatorType": "author"
					}
				],
				"date": "2015",
				"abstractNote": "黄瓜(Cucumis sativus L.)是我国最大的保护地栽培蔬菜作物,也是植物性别发育和维管束运输研究的重要模式植物。黄瓜基因组序列图谱已经构建完成,并且在此基础上又完成了全基因组SSR标记开发和涵盖330万个变异位点变异组图谱,成为黄瓜功能基因研究的重要平台和工具,相关转录组研究也有很多报道,不过共表达网络研究还是空白。本实验以温室型黄瓜9930为研究对象,选取10个不同组织,进行转录组测序,获得10份转录组原始数据。在对原始数据去除接头与低质量读段后,将高质量读段用Tophat2回贴到已经发表的栽培黄瓜基因组序列上。用Cufflinks对回贴后的数据计算FPKM值,获得10份组织的2...",
				"language": "中文;",
				"libraryCatalog": "CNKI",
				"thesisType": "硕士",
				"university": "南京农业大学",
				"url": "https://kns.cnki.net/KCMS/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201701&filename=1017045605.nh&v=MDc3ODZPZVorVnZGQ3ZrV3JyT1ZGMjZHYk84RzlmTXFwRWJQSVI4ZVgxTHV4WVM3RGgxVDNxVHJXTTFGckNVUkw=",
				"attachments": [
					{
						"title": "黄瓜共表达基因模块的识别及其特点分析",
						"mimeType": "text/html",
						"snapshot": true
					}
				],
				"tags": [
					{
						"tag": "co-expression"
					},
					{
						"tag": "cucumber"
					},
					{
						"tag": "network"
					},
					{
						"tag": "transcriptome"
					},
					{
						"tag": "共表达"
					},
					{
						"tag": "网络"
					},
					{
						"tag": "转录组"
					},
					{
						"tag": "黄瓜"
					}
				],
				"notes": [],
				"seeAlso": []
			}
		]
	}
]
/** END TEST CASES **/