Zotero Translator for the World Shakespeare Bibliography Online

World Shakespeare Bibliography Online.js

{
	"translatorID": "bf6b49e3-9198-4fbc-a559-a81fcfcce908",
	"label": "World Shakespeare Bibliography Online",
	"creator": "Matthias Heim",
	"target": "worldshakesbib.org",
	"minVersion": "1.0",
	"maxVersion": "",
	"priority": 100,
	"inRepository": true,
	"translatorType": 4,
	"browserSupport": "g",
	"lastUpdated": "2011-11-08 10:59:21"
}

/*
* Translator for the World Shakespeare Bibliography Online http://www.worldshakesbib.org
* Version 0.1: 3 November 2011 detectWeb complete
* Version 0.2: 4 November 2011 doWeb
* Version 0.4: 6 November 2011 initial version, fully working, but somewhat messy code with ugly workarounds
* Currently it works only for individual items, and collections of items saved by the user
* Currently, it does not offer Item Selection Dialogues for browse or search pages
* The World Shakespeare Bibliography Online offers extensive links to related documents,
* these are currently added as attached URLs to the corresponding entry in the bibliography,
* not as related items in Zotero, as translators can only add, but not access items in the 
* Zotero database.
* Written by Matthias Heim [email protected]
*/

var entries=new Array();

function Entry() {
	this.item_type="";
	this.dictionary=new Array();
	this.baseURL="";
	
	// returns one-dimensional array of table, uneven entries being the left side, even entries the right
	this.tableIntoArray = function(tableHTML) {
		currentTDstart=0;
		var TD=new Array();
	
		while ((currentTDstart=tableHTML.indexOf("<td>",currentTDstart))!=-1) {
			currentTD=tableHTML.substring(currentTDstart+4, (currentTDstart=tableHTML.indexOf("</td>",currentTDstart)));
		
			// remove enclosing <b> tag, enclosing <i> tag, enclosing brackets, and enclosing quotation marks
			if (currentTD.substr(0,3)+currentTD.substr(-4)=="<b></b>") currentTD=currentTD.slice(3,-4);
			if (currentTD.substr(0,3)+currentTD.substr(-4)=="<i></i>") {
				currentTD=currentTD.slice(3,-4);
				// check whether it really was enclosing, for e.g. 
				// currentTD="<i>Hamlet</i> to <i>The Tempest</i>";
				if (currentTD.indexOf("<i>")>currentTD.indexOf("</i>")) currentTD="<i>"+currentTD+"</i>";
			}
			if (currentTD.substr(0,1)+currentTD.substr(-1)=="[]") currentTD=currentTD.slice(1,-1);
			if (currentTD.substr(0,1)+currentTD.substr(-1)=="\"\"") currentTD=currentTD.slice(1,-1);
			
			// remove enclosing <a> tag, except if preceding element is a "Head Entry", in which case only the href-URL is retained
			if (currentTD.substr(0,2)+currentTD.substr(-4)=="<a</a>") {
				if (TD[TD.length-1]=="Head Entry") currentTD=currentTD.substring(currentTD.indexOf("href=\"")+6,currentTD.indexOf("\">"))
				else currentTD=currentTD.slice(currentTD.indexOf(">")+1,-4);
			}
		
			// remove font-color red tags, as these only highlight search items. Could probably be done with a regexp.
			while ((redstart=currentTD.indexOf('<font color="red">'))!=-1) {
				currentTD=currentTD.substring(0,redstart)+currentTD.substring(redstart+18);
				if ((redstart=currentTD.indexOf('</font>'))!=-1)
				 	currentTD=currentTD.substring(0,redstart)+currentTD.substring(redstart+7);
			}
		
		
			// remove final period? Yes, but watch out, it removes it also after initials
			if (currentTD.substr(-1)==".") currentTD=currentTD.slice(0,-1);
			if (currentTD.substr(-1)==",") currentTD=currentTD.slice(0,-1); // the case with many name entries
		
			//Zotero.debug(currentTD+"\n");
			TD.push(currentTD);
		}
	
		return TD;
	}

	this.addTable = function(tableHTML) {
		// turn table into array
		var tableArray=this.tableIntoArray(tableHTML);
		
		// Some names are followed by an entry for their roles, simply replace name-index entry with role
		if ((entryno=arrayTableIndexOf(tableArray,"Role"))!=-1) {
			tableArray[entryno-3]=tableArray[entryno];
			tableArray.splice(entryno-1,2);
		}
	
		var headEntry="";
		// If a head-entry exists, it needs to be extracted and later followed up!
		if ((entryno=arrayTableIndexOf(tableArray, "Head Entry"))!=-1) {
			headEntry=tableArray[entryno];
			tableArray.splice(entryno-1,2);
		}
		
		entryno=arrayTableIndexOf(tableArray, "Document Type");
		if (this.item_type=="") { // new item
			this.item_type=tableArray[entryno];
			tableArray.splice(entryno-1,2); // remove entry for document type
		} else {
			// check whether item_type has changed
			if (this.item_type!=tableArray[entryno]) {
				if ((this.item_type=="Article") && ((tableArray[entryno]=="Book monograph") || (tableArray[entryno]=="Book collection")) ) {
					this.item_type="bookSection"; // and not "Article"!
					tableArray.splice(entryno-1,2); // remove entry for document type
					
					tableArray[arrayTableIndexOf(tableArray, "Title")-1]="bookTitle";
					
					redundantname=arrayTableIndexOf(tableArray,"Name");
					if (redundantname!=-1) tableArray.splice(redundantname-1,2);
					
					tableArray.splice(arrayTableIndexOf(tableArray,"Notes/Performers")-1,2); // would refer only to collection, not to essay
					tableArray.splice(arrayTableIndexOf(tableArray,"Language")-1,2);
					tableArray.splice(arrayTableIndexOf(tableArray,"Record Number")-1,2);
					
					entryno2=arrayTableIndexOf(this.dictionary, "Venue/Publisher");
					this.dictionary[entryno2-1]="Pages";
					this.dictionary[entryno2]=this.dictionary[entryno2].substring(this.dictionary[entryno2].lastIndexOf(" "));
				} else {
					Zotero.debug("Unhandled change of item type exception. Item not saved.");
					// possibly a commit of the item to the database could be possible, but this event should actually never occur!
					return;
				}
			}
			// always retain the Index Location where the user found the document, hence discard it in follow-up entries
			arrayTableExtractItem(tableArray, "Index Location");
			arrayTableExtractItem(tableArray, "Record Number");
		}
		
		// assign all items from tableArray to this.dictionary
		for (entryno=0; entryno<tableArray.length; entryno+=2) {
			if ((entryno2=arrayTableIndexOf(this.dictionary, tableArray[entryno]))!=-1) {
				this.dictionary[entryno2]=tableArray[entryno+1];
			} else {
				this.dictionary.push(tableArray[entryno]);
				this.dictionary.push(tableArray[entryno+1]);
			}
		}
		
		if (headEntry!="") {
			// fetch next entry, follow up on item refered to in the table as Head Entry
			// retrieve item number (this) in global array, for later reference in callback
			for (var entrynum=0; entrynum<entries.length; entrynum++) if (entries[entrynum]===this) break;
			Zotero.Utilities.processDocuments(headEntry, function(newDoc) {entries[entrynum].addPage(newDoc)});
		} else {
			// commit to  ZoteroDatabase
			/*s="";
			for (dummy=0; dummy<this.dictionary.length; dummy+=2) s+=this.dictionary[dummy]+" : "+this.dictionary[dummy+1]+"\n";
			Zotero.debug(s);*/
			this.commitToZotero();
		}
	}
	
	this.commitToZotero = function() {
		switch (this.item_type) {
				case "Article":
					this.item_type = "journalArticle";
					break;
				case "Book monograph":
				case "Book collection": // Zotero does not distinguish the two
					this.item_type = "book"
					break;
				case "Dissertation":
					this.item_type = "thesis"
					break;
				case "Production":
					// this is most likely a theatre production, but videoRecording offers the closest alternative in Zotero
					this.item_type = "videoRecording";
					break;
				case "Audio Recording":
					this.item_type = "audioRecording";
					break;
				case "Film":
					this.item_type = "film";
					break;
				case "bookSection": break;
				default:
					Zotero.debug("This document type, "+this.item_type+", does not seem to exist in Zotero.")
					this.item_type = "";
					// unrecognized item? return empty, as no entry present
			}
		if (this.item_type=="") return; // unrecognized item
		Zotero.debug(this.item_type)
		var newItem = new Zotero.Item(this.item_type);
		newItem.title = arrayTableExtractItem(this.dictionary, "Title");
		
		if (this.item_type=="bookSection") {
			newItem.pages = arrayTableExtractItem(this.dictionary, "Pages");
			newItem.bookTitle = arrayTableExtractItem(this.dictionary, "bookTitle");
		}
		
		// standardize name entries
		for (counter=0; counter<this.dictionary.length; counter+=2) {
			switch (this.dictionary[counter]){
				case "Names":
				case "Name":
					this.dictionary[counter]="author";
					break;
				case "editor":
				case "editors":
					this.dictionary[counter]="editor";
					break;
				case "director":
				case "directors":
				case "conductor":
				case "conductors":
					this.dictionary[counter]="director";
					break;
				case "translators":
					this.dictionary[counter]="translator";
					break;
				case "narrator":
				case "narrators":
				case "lecturer":
				case "lecturers":
					this.dictionary[counter]="performer";
					break;
				case "general editor":
				case "general editors":
					this.dictionary[counter]="seriesEditor";
					break;
			}
		}
		
		// add names in order found on/in pages
		counter=0;
		while (counter<this.dictionary.length) {
			if ((this.dictionary[counter]=="director") || 
			    (this.dictionary[counter]=="performer") || 
			    (this.dictionary[counter]=="author") || 
			    (this.dictionary[counter]=="editor") || 
			    (this.dictionary[counter]=="translator") ||
			    (this.dictionary[counter]=="seriesEditor")) 
			{
				names=this.dictionary[counter+1].split(";");
				for (counter2=0; counter2<names.length; counter2++)
					newItem.creators.push(Zotero.Utilities.cleanAuthor(names[counter2], this.dictionary[counter], true));
				this.dictionary.splice(counter,2);
			} else counter+=2;
		}
		
		publisher=arrayTableExtractItem(this.dictionary, "Venue/Publisher");
		if (publisher!=-1) {
			// extract URL (from the rare items when it is present here, mostly obscure web-journals, no snapshot, as these URLs are often invalid)
			// again, this is only almost perfect. If a bookSection contains an URL, the pages will not be extracted correctly, and the 
			// (other) URL from the book's head entry will be taken into account. This is a a very rare case though.
			if ((url_index=publisher.indexOf("(http"))!=-1) { // common format
				url_lastindex=publisher.indexOf(")",url_index+4);
				if (url_lastindex==-1) url_lastindex=publisher.length;
				newItem.url=publisher.slice(url_index+1,url_lastindex);
				publisher=publisher.slice(0,url_index)+publisher.slice(url_lastindex+1);
				while (publisher.charAt(publisher.length-1)==" ") publisher=publisher.slice(0,publisher.length-1)
			}
			if ((url_index=publisher.indexOf("<a"))!=-1) { // common format for precise links
				if ((url_index=publisher.indexOf("href=",url_index))!=-1){
					url_lastindex=publisher.indexOf('"',url_index+6);
					if (url_lastindex==-1) url_lastindex=publisher.length; // an unlikely case
					newItem.url=publisher.slice(url_index+6,url_lastindex);
					publisher=publisher.slice(0,publisher.indexOf("<a"))+publisher.slice(publisher.indexOf("</a>")+4);
					while (publisher.charAt(publisher.length-1)==" ") publisher=publisher.slice(0,publisher.length-1)
				}
			}
			if ((url_index=publisher.indexOf("http"))!=-1) { // final common format, more difficult to parse safely
				url_lastindex=publisher.indexOf(" ",url_index+4);
				if (url_lastindex==-1) url_lastindex=publisher.length;
				newItem.url=publisher.slice(url_index,url_lastindex);
				publisher=publisher.slice(0,url_index)+publisher.slice(url_lastindex+1);
				while (publisher.charAt(publisher.length-1)==" ") publisher=publisher.slice(0,publisher.length-1)
			}
			
			//e.g. Manchester and New York: Manchester University Press, 2003. x + 227 pp.
			if ((this.item_type=="book") || (this.item_type=="bookSection")) {
		  	 	if ((this.item_type=="book") && (publisher.slice(-3)==" pp")) newItem.numPages=publisher.slice(publisher.slice(0,-3).lastIndexOf(" ")+1,-3);
		  		newItem.place=publisher.slice(0,publisher.indexOf(":"));
		  		newItem.publisher=publisher.slice(publisher.indexOf(":")+2, publisher.lastIndexOf(","));
		  	}
		
			// article publication info: e.g.
			// <i>Shakespeare Quarterly</i> 61, no. 2 (2010): 56-77
			if ((this.item_type=="journalArticle") || (this.item_type=="thesis")) {
		  	  	publisher+=" ";
		  		// dissertations are poorly parsed, but this format works for 
			  	// most dissertations in the database
			  	if (this.item_type!="thesis") // usually unpaginated
			  		newItem.pages=publisher.slice((i=publisher.indexOf("): ")+3),publisher.indexOf(" ", i));
			  	else newItem.university=publisher.slice(publisher.lastIndexOf("(")+1, publisher.lastIndexOf(")"));
		  	
			  	newItem.publicationTitle=publisher.slice(publisher.indexOf("<i>")+3,publisher.indexOf("</i>"));
			  	publisher=publisher.slice(publisher.indexOf("</i>")+5,publisher.indexOf(" ("));
			  	if ((issuestart=publisher.indexOf(", no."))!=-1) {
			  		newItem.volume=publisher.slice(0,issuestart);
			  		newItem.issue=publisher.slice(issuestart+6);
			  	} else newItem.volume=publisher;		  
			}
		
			// All the items below lack a standard format
			// The information usually includes date, label, running time, etc.
			// It is copied into the field that matches the usual information most closely
			if ((this.item_type=="audioRecording") || (this.item_type=="videoRecording") || (this.item_type=="film")) {
		  		if (this.item_type=="audioRecording") newItem.label=publisher;
		  		if (this.item_type=="videoRecording") newItem.videoRecordingFormat=publisher;
		  		if (this.item_type=="film") newItem.distributor=publisher;
		  	}
		}

		
		series=arrayTableExtractItem(this.dictionary, "Series Statement");
		if ((series!=-1) && ((this.item_type=="book") || (this.item_type=="bookSection") || (this.item_type=="journalArticle"))) newItem.series=series;
		if ((series!=-1) && ((this.item_type=="audioRecording") || (this.item_type="videoRecording"))) newItem.seriesTitle=series;
		
		language=arrayTableExtractItem(this.dictionary, "Language");
		if (language!=-1) newItem.language=language;
		archiveLocation=arrayTableExtractItem(this.dictionary, "Index Location");
		if (archiveLocation!=-1) newItem.archiveLocation=archiveLocation;
		callNumber=arrayTableExtractItem(this.dictionary, "Record Number");
		if (callNumber!=-1) newItem.callNumber=callNumber;
		
		date=arrayTableExtractItem(this.dictionary, "Date");
		if (date!=-1) newItem.date=date;
		
		arrayTableExtractItem(this.dictionary, "Cross Reference"); // To be discarded
		
		AdditionalTitleInfo=arrayTableExtractItem(this.dictionary, "Additional Title Info");
		NotesPerformers=arrayTableExtractItem(this.dictionary,"Notes/Performers");
		newItem.abstractNote=((AdditionalTitleInfo!=-1)?(AdditionalTitleInfo+((NotesPerformers!=-1)?"\n":"")):"")+((NotesPerformers!=-1)?NotesPerformers:"");
		
		reviews=arrayTableExtractItem(this.dictionary,"Reviews");
		if (reviews!=-1) {
			newItem.notes.push({note:reviews});
		}
		
		// Extract tags
		tags=arrayTableExtractItem(this.dictionary,"Descriptive Terms");
		if (tags!=-1) {
			newItem.tags=newItem.tags.concat(tags.split("; "));
		}
		tags=arrayTableExtractItem(this.dictionary,"Persons");
		if (tags!=-1) {
			newItem.tags=newItem.tags.concat(tags.split("; "));
		}
		
		
		// see also links are saved as attached URLs
		seeAlso=arrayTableExtractItem(this.dictionary,"See Also");
		if (seeAlso!=-1) {
			seeAlso=seeAlso.split("<a");
			for (i=0; i<seeAlso.length; i++) {
				if ((j=seeAlso[i].indexOf("href="))!=-1) {
					newItem.attachments.push({url:this.baseURL+seeAlso[i].slice(j+6,seeAlso[i].indexOf('"',j+6)), title:"See also: "+seeAlso[i].slice(seeAlso[i].lastIndexOf("</a>")+5), mimeType: "text/html", snapshot: false});
				}
			}	
		}
		
		while (arrayTableExtractItem(this.dictionary,"Document Type")!=-1); // sometimes survives as a duplicate, discard
		
		otherInformation="Other Information:";
		for (i=0; i<this.dictionary.length; i+=2)
		  otherInformation+="\n"+this.dictionary[i]+": "+this.dictionary[i+1];
		if (otherInformation!="Other Information:") newItem.notes.push({note:otherInformation});
		
		newItem.complete();
	}

	this.addPage =  function(doc) {
		if (doc.getElementById("records")!= null) {
			var records_content=doc.getElementById("records").innerHTML;
			// the entry is always presented in a table beginning in the same string
			var entry_start=records_content.indexOf("<tbody><tr><td><b>Index Location</b></td>");
			if (entry_start != -1) {
				this.addTable(records_content.substring(entry_start, records_content.indexOf("</tbody>", entry_start)+8));
			} else Zotero.debug("No entry could be found on page")
		} else Zotero.debug("No result could be found");
	}
}



function doWeb(doc, url) {
	
	// doWeb has its own parser, duplicating addPage above somewhat, because it is the 
	// only page where multiple results are possible, taking this into account here 
	// speeds the parsing up.
	
	// records are always contained within a div with the id 'records'
	if (doc.getElementById("records")!= null) {
		var records_content=doc.getElementById("records").innerHTML;
		// each entry is always presented in a table beginning in the same string
		var entry_start=0;
 		while ((entry_start=records_content.indexOf("<tbody><tr><td><b>Index Location</b></td>",entry_start))!=-1) { // at least one entry present
		  entries.push(new Entry());
		  entries[entries.length-1].baseURL=url.substring(0, url.indexOf("/",7));
		  entries[entries.length-1].addTable(records_content.substring(entry_start, (entry_start=records_content.indexOf("</tbody>", entry_start)+8)));
		}
	} else Zotero.debug("No results could be found"); // This is where search and browse results could be parsed!
}


function arrayTableIndexOf(narray, nvalue) {
	for (var neach=0; neach<narray.length; neach=neach+2) {
		if (narray[neach]==nvalue) return neach+1;
	}
	return -1;
}

function arrayTableExtractItem(narray, nvalue) {
	neach=arrayTableIndexOf(narray, nvalue);
	if (neach==-1) {
		return -1;
	} else {
		ncontent=narray[neach];
		narray.splice(neach-1,2);
		return ncontent;
	}
}

// detectWeb is self-contained and does not use any other functions, all of which relate to doWeb!
function detectWeb(doc, url) {
	// records are always contained within a div with the id 'records'
	if (doc.getElementById("records")!= null) {
		var records_content=doc.getElementById("records").innerHTML;
		// each entry is always presented in a table beginning in the same string
 	   var first_entry=records_content.indexOf("<tbody><tr><td><b>Index Location</b></td>"); // at least one entry present
		if (first_entry != -1) { // at least one entry is present
			// a note on "multiple"
			// This only works on the "View Saved Entries" (http://www.worldshakesbib.org/export) page, not in the search!
			// Completely different code would be necessary for browse or search pages
			// Since this page always already represents a selection made by the user, the handler will indiscriminately save all items, and not offer an Item Selection Dialogue
			// TODO: either behaviour may of course be changed in subsequent versions
			
			if (records_content.indexOf("<tbody><tr><td><b>Index Location</b></td>",first_entry+1)!=-1)	return "multiple"; // several entries present
		
			// if only one entry is present, its type can be retrieved from the 'Document type' entry in the table
			// e.g. <tr><td><b>Document Type</b></td><td>Article</td></tr>
			// Note that Article can also mean bookSection, a distinction that only the doWeb function will test, as it involves a GET command
			document_type=records_content.substring(startindex=(records_content.indexOf("<tr><td><b>Document Type</b></td><td>")+37),records_content.indexOf("</td>",startindex));
			switch (document_type) {
				case "Article":
					return "journalArticle"; // but could equally be "bookSection", see above
					break;
				case "Book monograph":
				case "Book collection": // Zotero does not distinguish the two
					return "book"
					break;
				case "Dissertation":
					return "thesis"
					break;
				case "Production":
					// this is most likely a theatre production, but videoRecording offers the closest alternative in Zotero
					return "videoRecording";
					break;
				case "Audio Recording":
					return "audioRecording";
					break;
				case "Film":
					return "film";
					break;
				default:
					// unrecognized item? return empty, as no entry present
			}
		} // else no entry present
	}
}