fix linter issues

Author: inhumantsar

.eslintrc.js

@@ -3,7 +3,7 @@
 
 module.exports = {
   "parserOptions": {
-    "ecmaVersion": 6,
+    "ecmaVersion": 2017,
   },
   "env": {
     "es6": true,

Readability.js

@@ -495,7 +495,7 @@ Readability.prototype = {
       // could assume it's the full title.
       var headings = this._concatNodeLists(
         doc.getElementsByTagName("h1"),
-        doc.getElementsByTagName("h2")
+        doc.getElementsByTagName("h2"),
       );
       var trimmedTitle = curTitle.trim();
       var match = this._someNode(headings, function(heading) {
@@ -1401,7 +1401,7 @@ Readability.prototype = {
           if (!parsed["@type"] && Array.isArray(parsed["@graph"])) {
             parsed = parsed["@graph"].find(function(it) {
               return (it["@type"] || "").match(
-                this.REGEXPS.jsonLdArticleTypes
+                this.REGEXPS.jsonLdArticleTypes,
               );
             });
           }
@@ -1470,6 +1470,36 @@ Readability.prototype = {
     return metadata ? metadata : {};
   },
 
+  /**
+   * Swaps the "Surname, GivenName" formatted bylines to "GivenName Surname".
+   *
+   * @param {string|string[]} name
+   * @returns Name or names in "GivenName Surname" format
+   */
+  _normalizeByline: function(name) {
+    var result = name;
+
+    if (Array.isArray(name)) {
+      return name.map((n) => this._normalizeByline(n));
+    }
+
+    // handle Surname, GivenName formatting
+    if (name.includes(",")) {
+      const parts = name.split(",").map(part => part.trim());
+      if (parts.length == 2) {
+        result = `${parts[1]} ${parts[0]}`;
+      }
+      if (parts.length > 2) {
+        result = `${parts[1]} ${parts[0]} ${parts.slice(2).join(" ")}`;
+      }
+    }
+
+    // remove things like "By:"
+    result = result.replace(/\w+:/, "");
+
+    return this._unescapeHtmlEntities(result);
+  },
+
   /**
    * Attempts to get excerpt and byline metadata for the article.
    *
@@ -1499,6 +1529,7 @@ Readability.prototype = {
       }
       var matches = null;
       var name = null;
+      var result = null;
 
       if (elementProperty) {
         matches = elementProperty.match(propertyPattern);
@@ -1507,7 +1538,7 @@ Readability.prototype = {
           // so we can match below.
           name = matches[0].toLowerCase().replace(/\s/g, "");
           // multiple authors
-          values[name] = content.trim();
+          result = content.trim();
         }
       }
       if (!matches && elementName && namePattern.test(elementName)) {
@@ -1516,8 +1547,23 @@ Readability.prototype = {
           // Convert to lowercase, remove any whitespace, and convert dots
           // to colons so we can match below.
           name = name.toLowerCase().replace(/\s/g, "").replace(/\./g, ":");
-          values[name] = content.trim();
+          result = content.trim();
+        }
+      }
+
+      // handle properties which might have multiple distinct values, eg: citation_author
+      if (result) {
+        if (values[name]) {
+          if (Array.isArray(values[name]) && typeof result == "string") {
+            values[name].push(result);
+          }
+          if (typeof values[name] == "string") {
+            values[name] = [values[name], result];
+          }
+        } else {
+          values[name] = result;
         }
+        this.log(`found metadata: ${name}=${values[name]}`);
       }
     });
 
@@ -1569,7 +1615,7 @@ Readability.prototype = {
     // in many sites the meta value is escaped with HTML entities,
     // so here we need to unescape it
     metadata.title = this._unescapeHtmlEntities(metadata.title);
-    metadata.byline = this._unescapeHtmlEntities(metadata.byline);
+    metadata.byline = this._normalizeByline(metadata.byline);
     metadata.excerpt = this._unescapeHtmlEntities(metadata.excerpt);
     metadata.siteName = this._unescapeHtmlEntities(metadata.siteName);
     metadata.publishedTime = this._unescapeHtmlEntities(metadata.publishedTime);

test/generate-testcase.js

@@ -62,8 +62,9 @@ function getWithRedirects(url, cb) {
       console.log("HEADERS:", JSON.stringify(response.headers));
     }
 
-    if(response.statusCode > 300 && response.statusCode <= 303) {
-      if (debug) console.log("following redirect", response.headers.location);
+    if (response.statusCode > 300 && response.statusCode <= 303) {
+      if (debug)
+        console.log("following redirect", response.headers.location);
       await getWithRedirects(response.headers.location, cb);
     }
 
@@ -73,7 +74,8 @@ function getWithRedirects(url, cb) {
     response.on("data", (chunk) => rv += chunk);
 
     response.on("end", () => {
-      if (debug) console.log("End received");
+      if (debug)
+        console.log("End received");
       cb(rv);
     });
   });
@@ -192,4 +194,4 @@ if (process.argv[2] === "all") {
   });
 } else {
   generateTestcase(process.argv[2]);
-}
+}

test/test-pages/nature/expected-metadata.json

@@ -1,6 +1,9 @@
 {
   "title": "Worldwide divergence of values",
-  "byline": "Medvedev, Danila",
+  "byline": [
+    "Joshua Conrad Jackson",
+    "Danila Medvedev"
+  ],
   "dir": null,
   "lang": "en",
   "excerpt": "Social scientists have long debated the nature of cultural change in a modernizing and globalizing world. Some scholars predicted that national cultures would converge by adopting social values typical of Western democracies. Others predicted that cultural differences in values would persist or even increase over time. We test these competing predictions by analyzing survey data from 1981 to 2022 (n = 406,185) from 76 national cultures. We find evidence of global value divergence. Values emphasizing tolerance and self-expression have diverged most sharply, especially between high-income Western countries and the rest of the world. We also find that countries with similar per-capita GDP levels have held similar values over the last 40 years. Over time, however, geographic proximity has emerged as an increasingly strong correlate of value similarity, indicating that values have diverged globally but converged regionally. The authors test whether social values have become converged or diverged across national cultures over the last 40 years using a 76-country analysis of the World Values Survey. They show that values have diverged, especially between high-income Western countries and the rest of the world.",

test/test-pages/nature/expected.html

@@ -651,9 +651,6 @@ <h2 id="additional-information"> Additional information </h2>
                 <b>Publisher’s note</b> Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
             </p>
         </div>
-        <div id="Sec25-section" data-title="Supplementary information">
-            <h2 id="Sec25"> Supplementary information </h2>
-        </div>
         <div id="rightslink-section" data-title="Rights and permissions">
             <h2 id="rightslink"> Rights and permissions </h2>
             <div id="rightslink-content">

test/test-pages/ourworldindata/expected-metadata.json

@@ -1,6 +1,6 @@
 {
   "title": "Why do we need to know about progress if we are concerned about the world's largest problems?",
-  "byline": "By: Max Roser",
+  "byline": "Max Roser",
   "dir": null,
   "excerpt": "Why have we made it our mission to publish “research and data to make progress against the world’s largest problems”?",
   "siteName": "Our World in Data",

test/test-readability.js

@@ -289,7 +289,7 @@ describe("Readability API", function() {
     it("should use custom video regex sent as option", function() {
       var dom = new JSDOM(
         "<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc mollis leo lacus, vitae semper nisl ullamcorper ut.</p>" +
-        "<iframe src=\"https://mycustomdomain.com/some-embeds\"></iframe>"
+        "<iframe src=\"https://mycustomdomain.com/some-embeds\"></iframe>",
       );
       var expected_xhtml = "<div id=\"readability-page-1\" class=\"page\">" +
         "<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc mollis leo lacus, vitae semper nisl ullamcorper ut.</p>" +