_getArticleMetadata()

in Readability.js [1767:1873]

• 79 lines of code
• 35 McCabe index (conditional complexity)

  _getArticleMetadata(jsonld) {
    var metadata = {};
    var values = {};
    var metaElements = this._doc.getElementsByTagName("meta");

    // property is a space-separated list of values
    var propertyPattern =
      /\s*(article|dc|dcterm|og|twitter)\s*:\s*(author|creator|description|published_time|title|site_name)\s*/gi;

    // name is a single value
    var namePattern =
      /^\s*(?:(dc|dcterm|og|twitter|parsely|weibo:(article|webpage))\s*[-\.:]\s*)?(author|creator|pub-date|description|title|site_name)\s*$/i;

    // Find description tags.
    this._forEachNode(metaElements, function (element) {
      var elementName = element.getAttribute("name");
      var elementProperty = element.getAttribute("property");
      var content = element.getAttribute("content");
      if (!content) {
        return;
      }
      var matches = null;
      var name = null;

      if (elementProperty) {
        matches = elementProperty.match(propertyPattern);
        if (matches) {
          // Convert to lowercase, and remove any whitespace
          // so we can match below.
          name = matches[0].toLowerCase().replace(/\s/g, "");
          // multiple authors
          values[name] = content.trim();
        }
      }
      if (!matches && elementName && namePattern.test(elementName)) {
        name = elementName;
        if (content) {
          // 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();
        }
      }
    });

    // get title
    metadata.title =
      jsonld.title ||
      values["dc:title"] ||
      values["dcterm:title"] ||
      values["og:title"] ||
      values["weibo:article:title"] ||
      values["weibo:webpage:title"] ||
      values.title ||
      values["twitter:title"] ||
      values["parsely-title"];

    if (!metadata.title) {
      metadata.title = this._getArticleTitle();
    }

    const articleAuthor =
      typeof values["article:author"] === "string" &&
      !this._isUrl(values["article:author"])
        ? values["article:author"]
        : undefined;

    // get author
    metadata.byline =
      jsonld.byline ||
      values["dc:creator"] ||
      values["dcterm:creator"] ||
      values.author ||
      values["parsely-author"] ||
      articleAuthor;

    // get description
    metadata.excerpt =
      jsonld.excerpt ||
      values["dc:description"] ||
      values["dcterm:description"] ||
      values["og:description"] ||
      values["weibo:article:description"] ||
      values["weibo:webpage:description"] ||
      values.description ||
      values["twitter:description"];

    // get site name
    metadata.siteName = jsonld.siteName || values["og:site_name"];

    // get article published time
    metadata.publishedTime =
      jsonld.datePublished ||
      values["article:published_time"] ||
      values["parsely-pub-date"] ||
      null;

    // 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.excerpt = this._unescapeHtmlEntities(metadata.excerpt);
    metadata.siteName = this._unescapeHtmlEntities(metadata.siteName);
    metadata.publishedTime = this._unescapeHtmlEntities(metadata.publishedTime);

    return metadata;
  },