A molecule link for org-mode
Posted March 26, 2016 at 03:28 PM | categories: orgmode, emacs, chemistry | tags:
Updated March 26, 2016 at 03:33 PM
Table of Contents
Here I am exploring some ideas on compact and functional representations of molecules in org-mode. We will use some functionality from OpenBabel (https://openbabel.org/docs/dev/index.html ) for conversion of formats.
One approach we could use is the SMILES representation. OpenBabel provides tools to convert SMILES to a visualization like this. Let's check out an old favorite: caffeine.
obabel -:"Cn1cnc2n(C)c(=O)n(C)c(=O)c12" -osvg
We can imagine the SMILES string is a program, and use an org-mode src block to contain it. It isn't quite a program, as it is more like data, but we can make the block executable if we define how to "execute" the block, and for that we will just have obabel generate the svg representation of the molecule. Here is our execute function. It simply generates the svg to stdout. We can use a :file header to capture it in a file.
(defun org-babel-execute:smiles (body params) (shell-command-to-string (format "obabel -:\"%s\" -osvg 2> /dev/null" body)))
You can find a smiles block in Appendix of molecules that was adapted from here .
Now, we need a link to refer to our molecule. We want the follow action to jump to our src block which should have a name. We will have it export as the name of the block linked to the molecule definition. This should work fine for definitions in the document. It is not robust to link to molecules in other org-files in the export. That would require those files to be exported too. For now we just define an HTML export.
(defun molecule-jump (name) (org-mark-ring-push) (org-open-link-from-string (format "[[%s]]" path))) (defun molecule-export (path desc backend) (let ((name (save-window-excursion (molecule-jump path) (org-element-property :name (org-element-context))))) (cond ((eq 'html backend) (format "<a href=\"#%s\">%s</a>" name name))))) (org-add-link-type "molecule" 'molecule-jump 'molecule-export)
Now we link to LSD and ethanol that allows us to navigate to the definition. We can also refer to a molecule in another file like ethanol. The links are clickable, and should jump to the molecule definition. On export to HTML they will be links to the definition.
Our link provides some limited functionality. We can provide more by making the follow action open a menu for example. Instead, we created a major mode here. It provides a function to convert smiles to CML. It is readily extensible to do other conversions.
One of the reasons we want to have molecules as "data" is so we can find them in our papers. Here is an example of that. We defined two molecules in the Appendix, and we find them here.
(org-element-map (org-element-parse-buffer) 'src-block (lambda (src) (when (string= "smiles" (org-element-property :language src)) (org-element-property :name src))))
There is still a lot to do to make this really functional. For example, we might want to use the molecules to write reactions. We might also want more advanced conversion or lookup functions, and more export options. It might be desirable to have tooltips on the links to see the molecules too. No doubt one might want to fine-tune the way the blocks run, so that options could be passed as header args. Maybe I will work on that another day.
1 Appendix of molecules
Here is an example smiles block.
2 smiles major mode
It would be nice to have a language mode to do special edits of SMILES src blocks. This mode does very little but provide a function that converts SMILES to CML using obabel and open it in a buffer. We redirect stderr to /dev/null to avoid seeing the messages from obabel. We also provide another function that opens a browser to names of the molecule.
(require 'easymenu) (defun smiles-cml () "Convert the smiles string in the buffer to CML." (interactive) (let ((smiles (buffer-string))) (switch-to-buffer (get-buffer-create "SMILES-CML")) (erase-buffer) (insert (shell-command-to-string (format "obabel -:\"%s\" -ocml 2> /dev/null" smiles))) (goto-char (point-min)) (xml-mode))) (defun smiles-names () (interactive) (browse-url (format "http://cactus.nci.nih.gov/chemical/structure/%s/names" (buffer-string)))) (defvar smiles-mode-map nil "Keymap for smiles-mode.") ;; adapted from http://ergoemacs.org/emacs/elisp_menu_for_major_mode.html (define-derived-mode smiles-mode fundamental-mode "smiles-mode" "Major mode for SMILES code." (setq buffer-invisibility-spec '(t) mode-name " ☺") (when (not smiles-mode-map) (setq smiles-mode-map (make-sparse-keymap))) (define-key smiles-mode-map (kbd "C-c C-c") 'smiles-cml) (define-key smiles-mode-map (kbd "C-c C-n") 'smiles-names) (define-key smiles-mode-map [menu-bar] (make-sparse-keymap)) (let ((menuMap (make-sparse-keymap "SMILES"))) (define-key smiles-mode-map [menu-bar smiles] (cons "SMILES" menuMap)) (define-key menuMap [cml] '("CML" . smiles-cml)) (define-key menuMap [names] '("Names" . smiles-names))))
Copyright (C) 2016 by John Kitchin. See the License for information about copying.
Org-mode version = 8.2.10