indian-ext

This package provides extensions to the standard Emacs ind-util.el functions.

You can install it from Melpa, with M-x package-install indian-ext

This library is currently focused on providing conversion methods between Devanāgarī and its romanized transliterations, but additions for other South Asian scripts are very welcome (but check ind-util.el first).

It defines the following extra decode/encode functions:

• between Velthuis and Devanāgarī:
  • `indian-ext-dev-velthuis-encode-region’
  • `indian-ext-dev-velthuis-decode-region’
• between SLP1 and Devanāgarī:
  • `indian-ext-dev-slp1-encode-region’
  • `indian-ext-dev-slp1-decode-region’
• between IAST and Devanāgarī:
  • `indian-ext-dev-iast-encode-region’
  • `indian-ext-dev-iast-decode-region’

The IAST and Velthuis are not case sensitive.

It also defines an additional input method, sanskrit-iast (use with `M-x set-input-method’).

You can find other input methods here:

• Including Karoṣṭhī: http://stefanbaums.com/unicode/sanskrit.el
• Requires login: http://indica-et-buddhica.org/repositorium/software/emacs-utf8-input-framework

See also the thread about websites for transliteration on the INDOLOGY.info mailing list starting here (October 2023): https://list.indology.info/pipermail/indology/2023-October/058241.html

As a library

There are functions to use these extensions in emacs-lisp programs:

• between Velthuis and Devanāgarī:
  • `indian-ext-dev-velthuis-encode-string’
  • `indian-ext-dev-velthuis-decode-string’
• between SLP1 and Devanāgarī:
  • `indian-ext-dev-slp1-encode-string’
  • `indian-ext-dev-slp1-decode-string’
• between IAST and Devanāgarī:
  • `indian-ext-dev-iast-encode-string’
  • `indian-ext-dev-iast-decode-string’

Speed is an issue, however:

(benchmark-run
    (dotimes (x 500)
      (cl-assert
       (string=
        (indian-ext-dev-iast-encode-string "ख.")
        "kha."))))

;; (4.902170626 30 4.18154369399997) most of the time seems to be
;; spent in ‘regex-opt’, called by ‘indian-regexp-of-hashtbl-keys’


;; Once it’s been byte-compiled, it’s okay though:
(byte-compile 'indian-ext-dev-iast-encode-string)
(benchmark-run
    (dotimes (x 500)
      (cl-assert
       (string=
        (indian-ext-dev-iast-encode-string "ख.")
        "kha."))))

;; (0.019145428 0 0.0)

;; and for comparison:

(benchmark-run
    (dotimes (x 500)
      (with-temp-buffer
        (insert "ख.")
        (goto-char (point-min))
        (while (re-search-forward "ख" nil t)
          (replace-match "kha"))
        (cl-assert
         (string=
          (buffer-string)
          "kha.")))))

;; (0.01554602 0 0.0)


;; Interfacing to ind-util.el

(require 'ind-util)

(with-temp-buffer
  (insert
   (indian-ext-dev-iast-decode-string "na yuktaḥ!"))
  (indian-dev-itrans-v5-encode-region (point-min) (point-max))
  (buffer-string))

;; Not working:
;; ;; To Bengali

;; (defun hacks/indian-bng-itrans-v5-encode-region (from to)
;;   (interactive "r")
;;   (indian-translate-region
;;    from to indian-bng-itrans-v5-hash t))

;; (defun hacks/indian-bng-itrans-v5-decode-region (from to)
;;   (interactive "r")
;;   (indian-translate-region
;;    from to indian-bng-itrans-v5-hash f))

;; (with-temp-buffer
;;   ;; convert to devanāgarī
;;   (insert
;;    (indian-ext-dev-iast-decode-string "na yuktaḥ!"))
;;   ;; encode it to itrans-v5
;;   (indian-dev-itrans-v5-encode-region (point-min) (point-max))
;;   ;; and convert that to Bengali (not working?)
;;   (hacks/indian-bng-itrans-v5-decode-region (point-min) (point-max))
;;   (buffer-string))