Make siunitx.el automatically recognize values and unit of unformatted text

Question

I use siunitx to typeset units in LaTeX. This is quite painless when typesetting it directly because of Emacs's siunitx.el (see below).

However, sometimes I copy exercises or solutions containing units from the Web or other sources which are not formatted.

So I paste something like 50 m, 50m/s, 1.2 m or 1,2 m (german comma) into a file in Emacs. Now I want to be able to mark this, type C-c C-m \SI and make Emacs automatically recognize units and transform them to \SI{50}{m}, \SI{50}{m\per s}, \SI{1,2}{m} and so on.

Any ideas how to modify siunitx.el to achieve this?

;;; siunitx.el --- AUCTeX style file for Siunitx
(TeX-add-style-hook "siunitx"
              (function
               (lambda ()
             (TeX-add-symbols
              '("SI"       "Value" "Unit")
                      '("ang"      "Angle")
              ))))

Answer 1

Something, perhaps a tad more automatic (but in general it's hard to get the representation right, you'd probably need to adjust that for your specific use-cases).

(defun try-replace-units (begin end)
  (interactive "r")
  (save-excursion
    (goto-char begin)
    (while
        (re-search-forward
         "\\<\\([-+]?\\([0-9]+\\|[0-9]*[.,][0-9]+\\)\\(e[0-9]+\\)?\\)\\s-*\\([a-zA-Z/]+\\)" end)
      (replace-match
       (let ((num (replace-regexp-in-string
                   "\\." "," (match-string 1)))
             (unit (match-string 4)))
         (setf end (+ end 8))
         (format "\\SI{%s}{%s}"
                 num
                 (replace-regexp-in-string
                  "/" "\\per " (or unit "") nil t)))
       nil t))))

Answer 2

Instead of changing the existing LaTeX-arg-siunitx-unit function, which is designed for interactive use, I would define a new function to do this, to be inserted in your init file. Here is a start of a solution

(defun mg-LaTeX-siunitx-SI-region (beg end)
  (interactive "r")
  (goto-char end)
  (insert TeX-grcl)
  (goto-char beg)
  (insert (concat TeX-esc "SI{"))
  (skip-chars-forward "[0-9\.,]")
  (insert (concat TeX-grcl TeX-grop)))

It can be improved to automatically remove space between value and unit.