lisp/gnus/html2text.el

   1 ;;; html2text.el --- a simple html to plain text converter
   2
   3 ;; Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
   4
   5 ;; Author: Joakim Hove <hove@phys.ntnu.no>
   6
   7 ;; This file is part of GNU Emacs.
   8
   9 ;; GNU Emacs is free software; you can redistribute it and/or modify
  10 ;; it under the terms of the GNU General Public License as published by
  11 ;; the Free Software Foundation; either version 2, or (at your option)
  12 ;; any later version.
  13
  14 ;; GNU Emacs is distributed in the hope that it will be useful,
  15 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17 ;; GNU General Public License for more details.
  18
  19 ;; You should have received a copy of the GNU General Public License
  20 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  21 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  22 ;; Boston, MA 02110-1301, USA.
  23
  24 ;;; Commentary:
  25
  26 ;; These functions provide a simple way to wash/clean html infected
  27 ;; mails.  Definitely do not work in all cases, but some improvement
  28 ;; in readability is generally obtained.  Formatting is only done in
  29 ;; the buffer, so the next time you enter the article it will be
  30 ;; "re-htmlized".
  31 ;;
  32 ;; The main function is `html2text'.
  33
  34 ;;; Code:
  35
  36 ;;
  37 ;; <Global variables>
  38 ;;
  39
  40 (eval-when-compile
  41   (require 'cl))
  42
  43 (defvar html2text-format-single-element-list '(("hr" . html2text-clean-hr)))
  44
  45 (defvar html2text-replace-list
  46   '(("&nbsp;" . " ") ("&gt;" . ">") ("&lt;" . "<") ("&quot;" . "\"")
  47     ("&amp;" . "&") ("&apos;" . "'"))
  48   "The map of entity to text.
  49
  50 This is an alist were each element is a dotted pair consisting of an
  51 old string, and a replacement string.  This replacement is done by the
  52 function `html2text-substitute' which basically performs a
  53 `replace-string' operation for every element in the list.  This is
  54 completely verbatim - without any use of REGEXP.")
  55
  56 (defvar html2text-remove-tag-list
  57   '("html" "body" "p" "img" "dir" "head" "div" "br" "font" "title" "meta")
  58   "A list of removable tags.
  59
  60 This is a list of tags which should be removed, without any
  61 formatting.  Note that tags in the list are presented *without*
  62 any \"<\" or \">\".  All occurrences of a tag appearing in this
  63 list are removed, irrespective of whether it is a closing or
  64 opening tag, or if the tag has additional attributes.  The
  65 deletion is done by the function `html2text-remove-tags'.
  66
  67 For instance the text:
  68
  69 \"Here comes something <font size\"+3\" face=\"Helvetica\"> big </font>.\"
  70
  71 will be reduced to:
  72
  73 \"Here comes something big.\"
  74
  75 If this list contains the element \"font\".")
  76
  77 (defvar html2text-format-tag-list
  78   '(("b"          . html2text-clean-bold)
  79     ("strong"     . html2text-clean-bold)
  80     ("u"          . html2text-clean-underline)
  81     ("i"          . html2text-clean-italic)
  82     ("em"         . html2text-clean-italic)
  83     ("blockquote" . html2text-clean-blockquote)
  84     ("a"          . html2text-clean-anchor)
  85     ("ul"         . html2text-clean-ul)
  86     ("ol"         . html2text-clean-ol)
  87     ("dl"         . html2text-clean-dl)
  88     ("center"     . html2text-clean-center))
  89   "An alist of tags and processing functions.
  90
  91 This is an alist where each dotted pair consists of a tag, and then
  92 the name of a function to be called when this tag is found.  The
  93 function is called with the arguments p1, p2, p3 and p4. These are
  94 demontrated below:
  95
  96 \"<b> This is bold text </b>\"
  97  ^   ^                 ^    ^
  98  |   |                 |    |
  99 p1  p2                p3   p4
 100
 101 Then the called function will typically format the text somewhat and
 102 remove the tags.")
 103
 104 (defvar html2text-remove-tag-list2  '("li" "dt" "dd" "meta")
 105   "Another list of removable tags.
 106
 107 This is a list of tags which are removed similarly to the list
 108 `html2text-remove-tag-list' - but these tags are retained for the
 109 formatting, and then moved afterward.")
 110
 111 ;;
 112 ;; </Global variables>
 113 ;;
 114
 115 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 116 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 117
 118 ;;
 119 ;; <Utility functions>
 120 ;;
 121
 122
 123 (defun html2text-replace-string (from-string to-string min max)
 124   "Replace FROM-STRING with TO-STRING in region from MIN to MAX."
 125   (goto-char min)
 126   (let ((delta (- (string-width to-string) (string-width from-string)))
 127         (change 0))
 128     (while (search-forward from-string max t)
 129       (replace-match to-string)
 130       (setq change (+ change delta)))
 131     change))
 132
 133 ;;
 134 ;; </Utility functions>
 135 ;;
 136
 137 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 138 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 139
 140 ;;
 141 ;; <Functions related to attributes> i.e. <font size=+3>
 142 ;;
 143
 144 (defun html2text-attr-value (list attribute)
 145   "Get value of ATTRIBUTE from LIST."
 146   (nth 1 (assoc attribute list)))
 147
 148 (defun html2text-get-attr (p1 p2)
 149   (goto-char p1)
 150   (re-search-forward " +[^ ]" p2 t)
 151   (let* ((attr-string (buffer-substring-no-properties (1- (point)) (1- p2)))
 152          (tmp-list (split-string attr-string))
 153          (attr-list)
 154          (counter 0)
 155          (prev (car tmp-list))
 156          (this (nth 1 tmp-list))
 157          (next (nth 2 tmp-list))
 158          (index 1))
 159
 160     (cond
 161      ;; size=3
 162      ((string-match "[^ ]=[^ ]" prev)
 163       (let ((attr  (nth 0 (split-string prev "=")))
 164             (value (nth 1 (split-string prev "="))))
 165         (setq attr-list (cons (list attr value) attr-list))))
 166      ;; size= 3
 167      ((string-match "[^ ]=\\'" prev)
 168       (setq attr-list (cons (list (substring prev 0 -1) this) attr-list))))
 169
 170     (while (< index (length tmp-list))
 171       (cond
 172        ;; size=3
 173        ((string-match "[^ ]=[^ ]" this)
 174         (let ((attr  (nth 0 (split-string this "=")))
 175               (value (nth 1 (split-string this "="))))
 176           (setq attr-list (cons (list attr value) attr-list))))
 177        ;; size =3
 178        ((string-match "\\`=[^ ]" this)
 179         (setq attr-list (cons (list prev (substring this 1)) attr-list)))
 180        ;; size= 3
 181        ((string-match "[^ ]=\\'" this)
 182         (setq attr-list (cons (list (substring this 0 -1) next) attr-list)))
 183        ;; size = 3
 184        ((string= "=" this)
 185         (setq attr-list (cons (list prev next) attr-list))))
 186       (setq index (1+ index))
 187       (setq prev this)
 188       (setq this next)
 189       (setq next (nth (1+ index) tmp-list)))
 190     ;;
 191     ;; Tags with no accompanying "=" i.e. value=nil
 192     ;;
 193     (setq prev (car tmp-list))
 194     (setq this (nth 1 tmp-list))
 195     (setq next (nth 2 tmp-list))
 196     (setq index 1)
 197
 198     (when (and (not (string-match "=" prev))
 199                (not (string= (substring this 0 1) "=")))
 200       (setq attr-list (cons (list prev nil) attr-list)))
 201     (while (< index (1- (length tmp-list)))
 202       (when (and (not (string-match "=" this))
 203                  (not (or (string= (substring next 0 1) "=")
 204                           (string= (substring prev -1) "="))))
 205         (setq attr-list (cons (list this nil) attr-list)))
 206       (setq index (1+ index))
 207       (setq prev this)
 208       (setq this next)
 209       (setq next (nth (1+ index) tmp-list)))
 210
 211     (when (and this
 212                (not (string-match "=" this))
 213                (not (string= (substring prev -1) "=")))
 214       (setq attr-list (cons (list this nil) attr-list)))
 215     ;; return - value
 216     attr-list))
 217
 218 ;;
 219 ;; </Functions related to attributes>
 220 ;;
 221
 222 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 223 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 224
 225 ;;
 226 ;; <Functions to be called to format a tag-pair>
 227 ;;
 228 (defun html2text-clean-list-items (p1 p2 list-type)
 229   (goto-char p1)
 230   (let ((item-nr 0)
 231         (items   0))
 232     (while (re-search-forward "<li>" p2 t)
 233       (setq items (1+ items)))
 234     (goto-char p1)
 235     (while (< item-nr items)
 236       (setq item-nr (1+ item-nr))
 237       (re-search-forward "<li>" (point-max) t)
 238       (cond
 239        ((string= list-type "ul") (insert " o "))
 240        ((string= list-type "ol") (insert (format " %s: " item-nr)))
 241        (t (insert " x "))))))
 242
 243 (defun html2text-clean-dtdd (p1 p2)
 244   (goto-char p1)
 245   (let ((items   0)
 246         (item-nr 0))
 247     (while (re-search-forward "<dt>" p2 t)
 248       (setq items (1+ items)))
 249     (goto-char p1)
 250     (while (< item-nr items)
 251       (setq item-nr (1+ item-nr))
 252       (re-search-forward "<dt>\\([ ]*\\)" (point-max) t)
 253       (when (match-string 1)
 254         (delete-region (point) (- (point) (string-width (match-string 1)))))
 255       (let ((def-p1 (point))
 256             (def-p2 0))
 257         (re-search-forward "\\([ ]*\\)\\(</dt>\\|<dd>\\)" (point-max) t)
 258         (if (match-string 1)
 259             (progn
 260               (let* ((mw1 (string-width (match-string 1)))
 261                      (mw2 (string-width (match-string 2)))
 262                      (mw  (+ mw1 mw2)))
 263                 (goto-char (- (point) mw))
 264                 (delete-region (point) (+ (point) mw1))
 265                 (setq def-p2 (point))))
 266           (setq def-p2 (- (point) (string-width (match-string 2)))))
 267         (put-text-property def-p1 def-p2 'face 'bold)))))
 268
 269 (defun html2text-delete-tags (p1 p2 p3 p4)
 270   (delete-region p1 p2)
 271   (delete-region (- p3 (- p2 p1)) (- p4 (- p2 p1))))
 272
 273 (defun html2text-delete-single-tag (p1 p2)
 274   (delete-region p1 p2))
 275
 276 (defun html2text-clean-hr (p1 p2)
 277   (html2text-delete-single-tag p1 p2)
 278   (goto-char p1)
 279   (newline 1)
 280   (insert (make-string fill-column ?-)))
 281
 282 (defun html2text-clean-ul (p1 p2 p3 p4)
 283   (html2text-delete-tags p1 p2 p3 p4)
 284   (html2text-clean-list-items p1 (- p3 (- p1 p2)) "ul"))
 285
 286 (defun html2text-clean-ol (p1 p2 p3 p4)
 287   (html2text-delete-tags p1 p2 p3 p4)
 288   (html2text-clean-list-items p1 (- p3 (- p1 p2)) "ol"))
 289
 290 (defun html2text-clean-dl (p1 p2 p3 p4)
 291   (html2text-delete-tags p1 p2 p3 p4)
 292   (html2text-clean-dtdd p1 (- p3 (- p1 p2))))
 293
 294 (defun html2text-clean-center (p1 p2 p3 p4)
 295   (html2text-delete-tags p1 p2 p3 p4)
 296   (center-region p1 (- p3 (- p2 p1))))
 297
 298 (defun html2text-clean-bold (p1 p2 p3 p4)
 299   (put-text-property p2 p3 'face 'bold)
 300   (html2text-delete-tags p1 p2 p3 p4))
 301
 302 (defun html2text-clean-title (p1 p2 p3 p4)
 303   (put-text-property p2 p3 'face 'bold)
 304   (html2text-delete-tags p1 p2 p3 p4))
 305
 306 (defun html2text-clean-underline (p1 p2 p3 p4)
 307   (put-text-property p2 p3 'face 'underline)
 308   (html2text-delete-tags p1 p2 p3 p4))
 309
 310 (defun html2text-clean-italic (p1 p2 p3 p4)
 311   (put-text-property p2 p3 'face 'italic)
 312   (html2text-delete-tags p1 p2 p3 p4))
 313
 314 (defun html2text-clean-font (p1 p2 p3 p4)
 315   (html2text-delete-tags p1 p2 p3 p4))
 316
 317 (defun html2text-clean-blockquote (p1 p2 p3 p4)
 318   (html2text-delete-tags p1 p2 p3 p4))
 319
 320 (defun html2text-clean-anchor (p1 p2 p3 p4)
 321   ;; If someone can explain how to make the URL clickable I will surely
 322   ;; improve upon this.
 323   ;; Maybe `goto-addr.el' can be used here.
 324   (let* ((attr-list (html2text-get-attr p1 p2))
 325          (href (html2text-attr-value attr-list "href")))
 326     (delete-region p1 p4)
 327     (when href
 328       (goto-char p1)
 329       (insert (substring href 1 -1 ))
 330       (put-text-property p1 (point) 'face 'bold))))
 331
 332 ;;
 333 ;; </Functions to be called to format a tag-pair>
 334 ;;
 335
 336 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 337 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 338
 339 ;;
 340 ;; <Functions to be called to fix up paragraphs>
 341 ;;
 342
 343 (defun html2text-fix-paragraph (p1 p2)
 344   (goto-char p1)
 345   (let ((has-br-line)
 346         (refill-start)
 347         (refill-stop))
 348     (when (re-search-forward "<br>$" p2 t)
 349       (goto-char p1)
 350       (when (re-search-forward ".+[^<][^b][^r][^>]$" p2 t)
 351         (beginning-of-line)
 352         (setq refill-start (point))
 353         (goto-char p2)
 354         (re-search-backward ".+[^<][^b][^r][^>]$" refill-start t)
 355         (next-line 1)
 356         (end-of-line)
 357         ;; refill-stop should ideally be adjusted to
 358         ;; accomodate the "<br>" strings which are removed
 359         ;; between refill-start and refill-stop.  Can simply
 360         ;; be returned from my-replace-string
 361         (setq refill-stop (+ (point)
 362                              (html2text-replace-string
 363                               "<br>" ""
 364                               refill-start (point))))
 365         ;; (message "Point = %s  refill-stop = %s" (point) refill-stop)
 366         ;; (sleep-for 4)
 367         (fill-region refill-start refill-stop))))
 368   (html2text-replace-string "<br>" "" p1 p2))
 369
 370 ;;
 371 ;; This one is interactive ...
 372 ;;
 373 (defun html2text-fix-paragraphs ()
 374   "This _tries_ to fix up the paragraphs - this is done in quite a ad-hook
 375 fashion, quite close to pure guess-work. It does work in some cases though."
 376   (interactive)
 377   (goto-char (point-min))
 378   (while (re-search-forward "^<br>$" nil t)
 379     (delete-region (match-beginning 0) (match-end 0)))
 380   ;; Removing lonely <br> on a single line, if they are left intact we
 381   ;; dont have any paragraphs at all.
 382   (goto-char (point-min))
 383   (while (not (eobp))
 384     (let ((p1 (point)))
 385       (forward-paragraph 1)
 386       ;;(message "Kaller fix med p1=%s  p2=%s " p1 (1- (point))) (sleep-for 5)
 387       (html2text-fix-paragraph p1 (1- (point)))
 388       (goto-char p1)
 389       (when (not (eobp))
 390         (forward-paragraph 1)))))
 391
 392 ;;
 393 ;; </Functions to be called to fix up paragraphs>
 394 ;;
 395
 396 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 397 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 398
 399 ;;
 400 ;; <Interactive functions>
 401 ;;
 402
 403 (defun html2text-remove-tags (tag-list)
 404   "Removes the tags listed in the list `html2text-remove-tag-list'.
 405 See the documentation for that variable."
 406   (interactive)
 407   (dolist (tag tag-list)
 408     (goto-char (point-min))
 409     (while (re-search-forward (format "\\(</?%s[^>]*>\\)" tag) (point-max) t)
 410       (delete-region (match-beginning 0) (match-end 0)))))
 411
 412 (defun html2text-format-tags ()
 413   "See the variable `html2text-format-tag-list' for documentation."
 414   (interactive)
 415   (dolist (tag-and-function html2text-format-tag-list)
 416     (let ((tag      (car tag-and-function))
 417           (function (cdr tag-and-function)))
 418       (goto-char (point-min))
 419       (while (re-search-forward (format "\\(<%s\\( [^>]*\\)?>\\)" tag)
 420                                 (point-max) t)
 421         (let ((p1)
 422               (p2 (point))
 423               (p3) (p4))
 424           (search-backward "<" (point-min) t)
 425           (setq p1 (point))
 426           (unless (search-forward (format "</%s>" tag) (point-max) t)
 427             (goto-char p2)
 428             (insert (format "</%s>" tag)))
 429           (setq p4 (point))
 430           (search-backward "</" (point-min) t)
 431           (setq p3 (point))
 432           (funcall function p1 p2 p3 p4)
 433           (goto-char p1))))))
 434
 435 (defun html2text-substitute ()
 436   "See the variable `html2text-replace-list' for documentation."
 437   (interactive)
 438   (dolist (e html2text-replace-list)
 439     (goto-char (point-min))
 440     (let ((old-string (car e))
 441           (new-string (cdr e)))
 442       (html2text-replace-string old-string new-string (point-min) (point-max)))))
 443
 444 (defun html2text-format-single-elements ()
 445   (interactive)
 446   (dolist (tag-and-function html2text-format-single-element-list)
 447     (let ((tag      (car tag-and-function))
 448           (function (cdr tag-and-function)))
 449       (goto-char (point-min))
 450       (while (re-search-forward (format "\\(<%s\\( [^>]*\\)?>\\)" tag)
 451                                 (point-max) t)
 452         (let ((p1)
 453               (p2 (point)))
 454           (search-backward "<" (point-min) t)
 455           (setq p1 (point))
 456           (funcall function p1 p2))))))
 457
 458 ;;
 459 ;; Main function
 460 ;;
 461
 462 ;;;###autoload
 463 (defun html2text ()
 464   "Convert HTML to plain text in the current buffer."
 465   (interactive)
 466   (save-excursion
 467     (let ((case-fold-search t)
 468           (buffer-read-only))
 469       (html2text-remove-tags html2text-remove-tag-list)
 470       (html2text-format-tags)
 471       (html2text-remove-tags html2text-remove-tag-list2)
 472       (html2text-substitute)
 473       (html2text-format-single-elements)
 474       (html2text-fix-paragraphs))))
 475
 476 ;;
 477 ;; </Interactive functions>
 478 ;;
 479 (provide 'html2text)
 480 ;;; arch-tag: e9e57b79-35d4-4de1-a647-e7e01fe56d1e
 481 ;;; html2text.el ends here