Refactoring of text-slicer

Now we process the rendered HTML of tiddlers, which allows us to
process HTML generated by MS Word. In fact, the HTML that MS Word
generates is so awful, I’ve instead been using Mammoth to do the
conversion: https://github.com/mwilliamson/mammoth.js

Also some necessary improvements to the fake dom implementation.

core/modules/utils/fakedom.js

@@ -43,6 +43,12 @@ var TW_Element = function(tag,namespace) {
 	this.namespaceURI = namespace || "http://www.w3.org/1999/xhtml";
 };
 
+Object.defineProperty(TW_Element.prototype, "nodeType", {
+	get: function() {
+		return 1;
+	}
+});
+
 TW_Element.prototype.setAttribute = function(name,value) {
 	if(this.isRaw) {
 		throw "Cannot setAttribute on a raw TW_Element";
@@ -93,6 +99,12 @@ TW_Element.prototype.hasChildNodes = function() {
 	return !!this.children.length;
 };
 
+Object.defineProperty(TW_Element.prototype, "childNodes", {
+	get: function() {
+		return this.children;
+	}
+});
+
 Object.defineProperty(TW_Element.prototype, "firstChild", {
 	get: function() {
 		return this.children[0];
@@ -103,6 +115,12 @@ TW_Element.prototype.addEventListener = function(type,listener,useCapture) {
 	// Do nothing
 };
 
+Object.defineProperty(TW_Element.prototype, "tagName", {
+	get: function() {
+		return this.tag || "";
+	}
+});
+
 Object.defineProperty(TW_Element.prototype, "className", {
 	get: function() {
 		return this.attributes["class"] || "";

plugins/tiddlywiki/text-slicer/slicer.js

@@ -23,21 +23,63 @@ var SLICER_OUTPUT_TITLE = "$:/TextSlicer";
 // Install the root widget event handlers
 exports.startup = function() {
 	$tw.rootWidget.addEventListener("tm-slice-tiddler",function(event) {
+		var slicer = new Slicer($tw.wiki,event.param);
+		// slicer.sliceTiddler();
+		// slicer.outputTiddlers();
 		// Slice up and output the tiddler
-		outputTiddlers(sliceTiddler(event.param),event.param,event.param);
+		slicer.outputTiddlers(slicer.sliceTiddler(event.param),event.param,event.param);
+		slicer.destroy();
 	});
 };
 
-var currentId = 0;
+function Slicer(wiki,sourceTitle) {
-
+	this.wiki = wiki;
-function nextId() {
+	this.sourceTitle = sourceTitle;
-	return ++currentId;
+	this.currentId = 0;
+	this.iframe = null; // Reference to iframe used for HTML parsing
 }
 
+Slicer.prototype.destroy = function() {
+	// Remove the iframe from the DOM
+	if(this.iframe && this.iframe.parentNode) {
+		this.iframe.parentNode.removeChild(this.iframe);
+	}
+};
+
+Slicer.prototype.nextId = function() {
+	return ++this.currentId;
+};
+
+Slicer.prototype.getSourceHtmlDocument = function(tiddler) {
+	this.iframe = document.createElement("iframe");
+	document.body.appendChild(this.iframe);
+	this.iframe.contentWindow.document.open();
+	this.iframe.contentWindow.document.write(tiddler.fields.text);
+	this.iframe.contentWindow.document.close();
+	return this.iframe.contentWindow.document;
+};
+
+Slicer.prototype.getSourceWikiDocument = function(tiddler) {
+	var widgetNode = this.wiki.makeTranscludeWidget(this.sourceTitle,{document: $tw.fakeDocument, parseAsInline: false}),
+		container = $tw.fakeDocument.createElement("div");
+	widgetNode.render(container,null);
+	return container;
+};
+
+Slicer.prototype.getSourceDocument = function() {
+	var tiddler = $tw.wiki.getTiddler(this.sourceTitle);
+	if(tiddler.fields.type === "text/html") {
+		return this.getSourceHtmlDocument(tiddler);
+	} else {
+		return this.getSourceWikiDocument(tiddler);
+	}
+};
+
 // Slice a tiddler into individual tiddlers
-function sliceTiddler(title) {
+Slicer.prototype.sliceTiddler = function(title) {
-	var tiddlers = {},
+	var self = this,
-		parser = $tw.wiki.parseTiddler(title),
+		tiddlers = {},
+		domNode = this.getSourceDocument(),
 		parentStack = [],
 		addTiddler = function(fields) {
 			if(fields.title) {
@@ -73,64 +115,83 @@ function sliceTiddler(title) {
 			} while(true);
 			return parentStack[parentStack.length - 1].title;
 		},
-		processNodeList = function(nodeList) {
+		isBlank = function(s) {
-			$tw.utils.each(nodeList,function(parseTreeNode) {
+			return (/^[\s\xA0]*$/mg).test(s);
+		},
+		processNodeList = function(domNodeList) {
+			$tw.utils.each(domNodeList,function(domNode) {
 				var parentTitle,
-					text = $tw.utils.getParseTreeText(parseTreeNode);
+					text = domNode.textContent,
-				if(parseTreeNode.type === "element" && (parseTreeNode.tag === "h1" || parseTreeNode.tag === "h2" || parseTreeNode.tag === "h3" || parseTreeNode.tag === "h4")) {
+					nodeType = domNode.nodeType;
-					parentTitle = popParentStackUntil(parseTreeNode.tag);
+				if(nodeType === 1) {
-					addToList(parentTitle,text);
+					var tagName = domNode.tagName.toLowerCase();
-					parentStack.push({type: parseTreeNode.tag, title: addTiddler({
+
-						title: text,
+					if(tagName === "p" && (domNode.getAttribute("style") || "").indexOf("mso-list:") !== -1) {
-						text: "<<display-heading-tiddler level:'" + parseTreeNode.tag + "'>>",
+						tagName = "li";
-						list: [],
+					}
-						tags: [parentTitle]
+
-					})});
+					if(tagName === "h1" || tagName === "h2" || tagName === "h3" || tagName === "h4") {
-				} else if(parseTreeNode.type === "element" && (parseTreeNode.tag === "ul" || parseTreeNode.tag === "ol")) {
+						if(!isBlank(text)) {
-					var listTitle = title + "-list-" + nextId();
+							parentTitle = popParentStackUntil(tagName);
-					parentTitle = parentStack[parentStack.length - 1].title;
+							addToList(parentTitle,text);
-					addToList(parentTitle,listTitle);
+							parentStack.push({type: tagName, title: addTiddler({
-					parentStack.push({type: parseTreeNode.tag, title: addTiddler({
+								title: text,
-						title: listTitle,
+								text: "<<display-heading-tiddler level:'" + tagName + "'>>",
-						text: "<<display-list-tiddler type:'" + parseTreeNode.tag + "'>>",
+								list: [],
-						list: [],
+								tags: [parentTitle]
-						tags: [parentTitle]
+							})});
-					})});
+						}
-					processNodeList(parseTreeNode.children);
+					} else if(tagName === "ul" || tagName === "ol") {
-					parentStack.pop();
+						var listTitle = title + "-list-" + self.nextId();
-				} else if(parseTreeNode.type === "element" && parseTreeNode.tag === "li") {
+						parentTitle = parentStack[parentStack.length - 1].title;
-					var listItemTitle = title + "-listitem-" + nextId();
+						addToList(parentTitle,listTitle);
-					parentTitle = parentStack[parentStack.length - 1].title;
+						parentStack.push({type: tagName, title: addTiddler({
-					addToList(parentTitle,listItemTitle);
+							title: listTitle,
-					addTiddler({
+							text: "<<display-list-tiddler type:'" + tagName + "'>>",
-						title: listItemTitle,
+							list: [],
-						text: text,
+							tags: [parentTitle]
-						list: [],
+						})});
-						tags: [parentTitle]
+						processNodeList(domNode.childNodes);
-					});
+						parentStack.pop();
-				} else if(parseTreeNode.type === "element" && parseTreeNode.tag === "p") {
+					} else if(tagName === "li") {
-					parentTitle = parentStack[parentStack.length - 1].title;
+						if(!isBlank(text)) {
-					addToList(parentTitle,addTiddler({
+							var listItemTitle = title + "-listitem-" + self.nextId();
-						title: title + "-para-" + nextId(),
+							parentTitle = parentStack[parentStack.length - 1].title;
-						text: text,
+							addToList(parentTitle,listItemTitle);
-						tags: [parentTitle]
+							addTiddler({
-					}));
+								title: listItemTitle,
+								text: text,
+								list: [],
+								tags: [parentTitle]
+							});
+						}
+					} else if(tagName === "p") {
+						if(!isBlank(text)) {
+							parentTitle = parentStack[parentStack.length - 1].title;
+							addToList(parentTitle,addTiddler({
+								title: title + "-para-" + self.nextId(),
+								text: text,
+								tags: [parentTitle]
+							}));
+						}
+					} else if(domNode.hasChildNodes()) {
+						processNodeList(domNode.childNodes);
+					}
 				}
 			});
 		};
-	if(parser) {
+	parentStack.push({type: "h0", title: addTiddler({
-		parentStack.push({type: "h0", title: addTiddler({
+		title: "Sliced up " + title,
-			title: "Sliced up " + title,
+		text: "{{||$:/plugins/tiddlywiki/text-slicer/templates/display-document}}",
-			text: "{{||$:/plugins/tiddlywiki/text-slicer/templates/display-document}}",
+		list: []
-			list: []
+	})});
-		})});
+console.log(domNode);
-		processNodeList(parser.tree);
+	processNodeList(domNode.childNodes);
-	}
 	return tiddlers;
-}
+};
 
 // Output directly to the output tiddlers
-function outputTiddlers(tiddlers,title,navigateFromTitle) {
+Slicer.prototype.outputTiddlers = function(tiddlers,title,navigateFromTitle) {
 	$tw.utils.each(tiddlers,function(tiddlerFields) {
 		var title = tiddlerFields.title;
 		if(title) {
@@ -140,10 +201,10 @@ function outputTiddlers(tiddlers,title,navigateFromTitle) {
 	// Navigate to output
 	var story = new $tw.Story({wiki: $tw.wiki});
 	story.navigateTiddler("Sliced up " + title,navigateFromTitle);
-}
+};
 
 // Output via an import tiddler
-function outputTiddlers_viaImportTiddler(tiddlers,navigateFromTitle) {
+Slicer.prototype.outputTiddlers_viaImportTiddler = function(tiddlers,navigateFromTitle) {
 	// Get the current slicer output tiddler
 	var slicerOutputTiddler = $tw.wiki.getTiddler(SLICER_OUTPUT_TITLE),
 		slicerOutputData = $tw.wiki.getTiddlerData(SLICER_OUTPUT_TITLE,{}),
@@ -167,6 +228,6 @@ function outputTiddlers_viaImportTiddler(tiddlers,navigateFromTitle) {
 	// Navigate to output
 	var story = new $tw.Story({wiki: $tw.wiki});
 	story.navigateTiddler(SLICER_OUTPUT_TITLE,navigateFromTitle);
-}
+};
 
 })();