upgrade readability

content.js

@@ -24,10 +24,12 @@ function pageTitle() {
 }
 
 function getReadableDocument() {
-	// Readability directly change the passed document so clone it so as
-	// to preserve the original web page.
+	// Readability directly change the passed document, so clone to preserve the original web page.
 	const documentCopy = document.cloneNode(true);
-	const readability = new Readability(documentCopy);
+	const readability = new Readability(documentCopy, {
+		serializer: el => el // so that .content is returned as DOM element instead of HTML
+	});
+
 	const article = readability.parse();
 
 	if (!article) {
@@ -36,7 +38,7 @@ function getReadableDocument() {
 
 	return {
 		title: article.title,
-		body: article.articleContent,
+		body: article.content,
 	}
 }
 

lib/JSDOMParser.js

@@ -278,7 +278,7 @@
 
   var whitespace = [" ", "\t", "\n", "\r"];
 
-  // See http://www.w3schools.com/dom/dom_nodetype.asp
+  // See https://developer.mozilla.org/en-US/docs/Web/API/Node/nodeType
   var nodeTypes = {
     ELEMENT_NODE: 1,
     ATTRIBUTE_NODE: 2,
@@ -705,7 +705,6 @@
       }
 
       // Using Array.join() avoids the overhead from lazy string concatenation.
-      // See http://blog.cdleary.com/2012/01/string-representation-in-spidermonkey/#ropes
       var arr = [];
       getHTML(this);
       return arr.join("");
@@ -875,7 +874,11 @@
 
   JSDOMParser.prototype = {
     error: function(m) {
-      dump("JSDOMParser error: " + m + "\n");
+      if (typeof dump !== "undefined") {
+        dump("JSDOMParser error: " + m + "\n");
+      } else if (typeof console !== "undefined") {
+        console.log("JSDOMParser error: " + m + "\n");
+      }
       this.errorState += m + "\n";
     },
 
@@ -1187,3 +1190,7 @@
   global.JSDOMParser = JSDOMParser;
 
 })(this);
+
+if (typeof module === "object") {
+  module.exports = this.JSDOMParser;
+}

lib/Readability-readerable.js

@@ -1,5 +1,4 @@
 /* eslint-env es6:false */
-/* globals exports */
 /*
  * Copyright (c) 2010 Arc90 Inc
  *
@@ -31,22 +30,30 @@ var REGEXPS = {
 function isNodeVisible(node) {
   // Have to null-check node.style and node.className.indexOf to deal with SVG and MathML nodes.
   return (!node.style || node.style.display != "none")
-      && !node.hasAttribute("hidden")
-      //check for "fallback-image" so that wikimedia math images are displayed
-      && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || (node.className && node.className.indexOf && node.className.indexOf("fallback-image") !== -1));
+    && !node.hasAttribute("hidden")
+    //check for "fallback-image" so that wikimedia math images are displayed
+    && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || (node.className && node.className.indexOf && node.className.indexOf("fallback-image") !== -1));
 }
 
 /**
  * Decides whether or not the document is reader-able without parsing the whole thing.
- *
- * @return boolean Whether or not we suspect Readability.parse() will suceeed at returning an article object.
+ * @param {Object} options Configuration object.
+ * @param {number} [options.minContentLength=140] The minimum node content length used to decide if the document is readerable.
+ * @param {number} [options.minScore=20] The minumum cumulated 'score' used to determine if the document is readerable.
+ * @param {Function} [options.visibilityChecker=isNodeVisible] The function used to determine if a node is visible.
+ * @return {boolean} Whether or not we suspect Readability.parse() will suceeed at returning an article object.
  */
-function isProbablyReaderable(doc, isVisible) {
-  if (!isVisible) {
-    isVisible = isNodeVisible;
+function isProbablyReaderable(doc, options = {}) {
+  // For backward compatibility reasons 'options' can either be a configuration object or the function used
+  // to determine if a node is visible.
+  if (typeof options == "function") {
+    options = { visibilityChecker: options };
   }
 
-  var nodes = doc.querySelectorAll("p, pre");
+  var defaultOptions = { minScore: 20, minContentLength: 140, visibilityChecker: isNodeVisible };
+  options = Object.assign(defaultOptions, options);
+
+  var nodes = doc.querySelectorAll("p, pre, article");
 
   // Get <div> nodes which have <br> node(s) and append them into the `nodes` variable.
   // Some articles' DOM structures might look like
@@ -58,7 +65,7 @@ function isProbablyReaderable(doc, isVisible) {
   var brNodes = doc.querySelectorAll("div > br");
   if (brNodes.length) {
     var set = new Set(nodes);
-    [].forEach.call(brNodes, function(node) {
+    [].forEach.call(brNodes, function (node) {
       set.add(node.parentNode);
     });
     nodes = Array.from(set);
@@ -67,9 +74,10 @@ function isProbablyReaderable(doc, isVisible) {
   var score = 0;
   // This is a little cheeky, we use the accumulator 'score' to decide what to return from
   // this callback:
-  return [].some.call(nodes, function(node) {
-    if (!isVisible(node))
+  return [].some.call(nodes, function (node) {
+    if (!options.visibilityChecker(node)) {
       return false;
+    }
 
     var matchString = node.className + " " + node.id;
     if (REGEXPS.unlikelyCandidates.test(matchString) &&
@@ -82,19 +90,19 @@ function isProbablyReaderable(doc, isVisible) {
     }
 
     var textContentLength = node.textContent.trim().length;
-    if (textContentLength < 140) {
+    if (textContentLength < options.minContentLength) {
       return false;
     }
 
-    score += Math.sqrt(textContentLength - 140);
+    score += Math.sqrt(textContentLength - options.minContentLength);
 
-    if (score > 20) {
+    if (score > options.minScore) {
       return true;
     }
     return false;
   });
 }
 
-if (typeof exports === "object") {
-  exports.isProbablyReaderable = isProbablyReaderable;
+if (typeof module === "object") {
+  module.exports = isProbablyReaderable;
 }