Text extraction after dom traversal returns empty

Question

I would like to extract the text from each visited node but while traversing and build a list. I only get an empty result. Here is the following attempt using dom.el

(defun extract-text-from-dom (dom-node)
  "Recursively extract text from a DOM-NODE."
  (cond
   ((null dom-node) "")
   ((dom-text dom-node) (dom-text dom-node))
   (t
    (mapconcat
     #'extract-text-from-dom
     (dom-children dom-node)
     " "))))

(let* ((html "<div><p>This is a <b>sample</b> text.</p> <br><h1>Another Text</h1></div>")
       (dom-tree (libxml-parse-html-region 1 (length html) html))
       (text (extract-text-from-dom dom-tree)))
  (message "Extracted Text: %s" text))

Answer 1

One thing that I notice is that libxml-parse-html-region doesn’t take a string to parse. Instead it parses a region of the current buffer. Use C-h f to look up help for functions. It will tell you what arguments they take, and the docstring will describe those arguments in detail. Try this instead:

(let ((html "<div><p>This is a <b>sample</b> text.</p> <br><h1>Another Text</h1></div>"))
  (with-temp-buffer
    (insert html)
    (let ((text (extract-text-from-dom (libxml-parse-html-region (point-min) (point-max)))))
      (message "Extracted Text: %s" text))))

Note also that dom nodes can have both text and children, so your extraction function isn’t quite right.