Suggestion: follow links in json-ld to find author name

[danielnixon]

Consider the json-ld from this article: https://harpers.org/archive/2024/12/the-painted-protest-dean-kissick-contemporary-art//

It looks something like this (truncated to just the bits relevant to the author name):

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Article",
      ...
      ...
      ...
      "author": [
        {
          "@id": "https://harpers.org/#/schema/person/69e72b14e23945c24e35bf38c40c0fd6"
        }
      ],
      ...
      ...
      ...
    },
    ...
    ...
    ...
    {
      "@type": "Person",
      "@id": "https://harpers.org/#/schema/person/69e72b14e23945c24e35bf38c40c0fd6",
      "name": "Dean Kissick",
      ...
      ...
      ...
    }
  ]
}

Readability will currently stop following the graph once it gets to the author element that doesn't have a name. If instead we followed the author ID to the person element, we'd find the author name there.

Because Readability doesn't see an author in the json-ld, it falls back on the rel="author" link present in the document body, which unfortunately contains "by Dean Kissick," with the extraneous text "by" and the trailing comma.

[danielnixon]

Quick and dirty patch for argument's sake:

             return;
           }
 
-          if (!parsed["@type"] && Array.isArray(parsed["@graph"])) {
-            parsed = parsed["@graph"].find(function(it) {
+          var graphArray = parsed["@graph"];
+          if (!parsed["@type"] && Array.isArray(graphArray)) {
+            parsed = graphArray.find(function(it) {
               return (it["@type"] || "").match(
                 this.REGEXPS.jsonLdArticleTypes
               );
@@ -1436,6 +1437,15 @@ Readability.prototype = {
                   return author.name.trim();
                 })
                 .join(", ");
+            } else if (typeof parsed.author["@id"] === "string") {
+              var authorId = parsed.author["@id"];
+              var authorPerson = graphArray.find(function(it) {
+                return (it["@id"] || "").match(authorId);
+              });
+              if (authorPerson && typeof authorPerson.name === "string") {
+                metadata.byline = authorPerson.name.trim();
+              }
             }
           }
           if (typeof parsed.description === "string") {