_prepArticle()

in Readability.js [792:894]
43 lines of code
2 McCabe index (conditional complexity)

  _prepArticle(articleContent) {
    this._cleanStyles(articleContent);

    // Check for data tables before we continue, to avoid removing items in
    // those tables, which will often be isolated even though they're
    // visually linked to other content-ful elements (text, images, etc.).
    this._markDataTables(articleContent);

    this._fixLazyImages(articleContent);

    // Clean out junk from the article content
    this._cleanConditionally(articleContent, "form");
    this._cleanConditionally(articleContent, "fieldset");
    this._clean(articleContent, "object");
    this._clean(articleContent, "embed");
    this._clean(articleContent, "footer");
    this._clean(articleContent, "link");
    this._clean(articleContent, "aside");

    // Clean out elements with little content that have "share" in their id/class combinations from final top candidates,
    // which means we don't remove the top candidates even they have "share".

    var shareElementThreshold = this.DEFAULT_CHAR_THRESHOLD;

    this._forEachNode(articleContent.children, function (topCandidate) {
      this._cleanMatchedNodes(topCandidate, function (node, matchString) {
        return (
          this.REGEXPS.shareElements.test(matchString) &&
          node.textContent.length < shareElementThreshold
        );
      });
    });

    this._clean(articleContent, "iframe");
    this._clean(articleContent, "input");
    this._clean(articleContent, "textarea");
    this._clean(articleContent, "select");
    this._clean(articleContent, "button");
    this._cleanHeaders(articleContent);

    // Do these last as the previous stuff may have removed junk
    // that will affect these
    this._cleanConditionally(articleContent, "table");
    this._cleanConditionally(articleContent, "ul");
    this._cleanConditionally(articleContent, "div");

    // replace H1 with H2 as H1 should be only title that is displayed separately
    this._replaceNodeTags(
      this._getAllNodesWithTag(articleContent, ["h1"]),
      "h2"
    );

    // Remove extra paragraphs
    this._removeNodes(
      this._getAllNodesWithTag(articleContent, ["p"]),
      function (paragraph) {
        // At this point, nasty iframes have been removed; only embedded video
        // ones remain.
        var contentElementCount = this._getAllNodesWithTag(paragraph, [
          "img",
          "embed",
          "object",
          "iframe",
        ]).length;
        return (
          contentElementCount === 0 && !this._getInnerText(paragraph, false)
        );
      }
    );

    this._forEachNode(
      this._getAllNodesWithTag(articleContent, ["br"]),
      function (br) {
        var next = this._nextNode(br.nextSibling);
        if (next && next.tagName == "P") {
          br.remove();
        }
      }
    );

    // Remove single-cell tables
    this._forEachNode(
      this._getAllNodesWithTag(articleContent, ["table"]),
      function (table) {
        var tbody = this._hasSingleTagInsideElement(table, "TBODY")
          ? table.firstElementChild
          : table;
        if (this._hasSingleTagInsideElement(tbody, "TR")) {
          var row = tbody.firstElementChild;
          if (this._hasSingleTagInsideElement(row, "TD")) {
            var cell = row.firstElementChild;
            cell = this._setNodeTag(
              cell,
              this._everyNode(cell.childNodes, this._isPhrasingContent)
                ? "P"
                : "DIV"
            );
            table.parentNode.replaceChild(cell, table);
          }
        }
      }
    );
  },