code.delx.au - gnu-emacs/blob - lisp/gnus/html2text.el

   1 ;;; html2text.el --- a simple html to plain text converter
   2 ;; Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
   3
   4 ;; Author: Joakim Hove <hove@phys.ntnu.no>
   5
   6 ;; This file is part of GNU Emacs.
   7
   8 ;; GNU Emacs is free software; you can redistribute it and/or modify
   9 ;; it under the terms of the GNU General Public License as published by
  10 ;; the Free Software Foundation; either version 2, or (at your option)
  11 ;; any later version.
  12
  13 ;; GNU Emacs is distributed in the hope that it will be useful,
  14 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 ;; GNU General Public License for more details.
  17
  18 ;; You should have received a copy of the GNU General Public License
  19 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  20 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  21 ;; Boston, MA 02111-1307, USA.
  22
  23 ;;; Commentary:
  24
  25 ;; These functions provide a simple way to wash/clean html infected
  26 ;; mails.  Definitely do not work in all cases, but some improvement
  27 ;; in readability is generally obtained. Formatting is only done in
  28 ;; the buffer, so the next time you enter the article it will be
  29 ;; "re-htmlized".
  30 ;;
  31 ;; The main function is "html2text"
  32
  33 ;;; Code:
  34
  35 ;;
  36 ;; <Global variables>
  37 ;;
  38
  39 (eval-when-compile
  40   (require 'cl))
  41
  42 (defvar html2text-format-single-element-list '(("hr" . html2text-clean-hr)))
  43
  44 (defvar html2text-replace-list
  45   '(("&nbsp;" . " ") ("&gt;" . ">") ("&lt;" . "<") ("&quot;" . "\"")
  46     ("&amp;" . "&") ("&apos;" . "'"))
  47   "The map of entity to text.
  48
  49 This is an alist were each element is a dotted pair consisting of an
  50 old string, and a replacement string. This replacement is done by the
  51 function \"html2text-substitute\" which basically performs a
  52 replace-string operation for every element in the list. This is
  53 completely verbatim - without any use of REGEXP.")
  54
  55 (defvar html2text-remove-tag-list
  56   '("html" "body" "p" "img" "dir" "head" "div" "br" "font" "title" "meta")
  57   "A list of removable tags.
  58
  59 This is a list of tags which should be removed, without any
  60 formatting.  Observe that if you the tags in the list are presented
  61 *without* any \"<\" or \">\". All occurences of a tag appearing in
  62 this list are removed, irrespective of whether it is a closing or
  63 opening tag, or if the tag has additional attributes. The actual
  64 deletion is done by the function \"html2text-remove-tags\".
  65
  66 For instance the text:
  67
  68 \"Here comes something <font size\"+3\" face=\"Helvetica\"> big </font>.\"
  69
  70 will be reduced to:
  71
  72 \"Here comes something big.\"
  73
  74 If this list contains the element \"font\".")
  75
  76 (defvar html2text-format-tag-list
  77   '(("b"          . html2text-clean-bold)
  78     ("u"          . html2text-clean-underline)
  79     ("i"          . html2text-clean-italic)
  80     ("blockquote" . html2text-clean-blockquote)
  81     ("a"          . html2text-clean-anchor)
  82     ("ul"         . html2text-clean-ul)
  83     ("ol"         . html2text-clean-ol)
  84     ("dl"         . html2text-clean-dl)
  85     ("center"     . html2text-clean-center))
  86   "An alist of tags and processing functions.
  87
  88 This is an alist where each dotted pair consists of a tag, and then
  89 the name of a function to be called when this tag is found. The
  90 function is called with the arguments p1, p2, p3 and p4. These are
  91 demontrated below:
  92
  93 \"<b> This is bold text </b>\"
  94  ^   ^                 ^    ^
  95  |   |                 |    |
  96 p1  p2                p3   p4
  97
  98 Then the called function will typically format the text somewhat and
  99 remove the tags.")
 100
 101 (defvar html2text-remove-tag-list2  '("li" "dt" "dd" "meta")
 102   "Another list of removable tags.
 103
 104 This is a list of tags which are removed similarly to the list
 105 `html2text-remove-tag-list' - but these tags are retained for the
 106 formatting, and then moved afterward.")
 107
 108 ;;
 109 ;; </Global variables>
 110 ;;
 111
 112 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 113 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 114
 115 ;;
 116 ;; <Utility functions>
 117 ;;
 118
 119
 120 (defun html2text-replace-string (from-string to-string p1 p2)
 121   (goto-char p1)
 122   (let ((delta (- (string-width to-string) (string-width from-string)))
 123         (change 0))
 124     (while (search-forward from-string p2 t)
 125       (replace-match to-string)
 126       (setq change (+ change delta))
 127       )
 128     change
 129     )
 130   )
 131
 132 ;;
 133 ;; </Utility functions>
 134 ;;
 135
 136 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 137 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 138
 139 ;;
 140 ;; <Functions related to attributes> i.e. <font size=+3>
 141 ;;
 142
 143 (defun html2text-attr-value (attr-list attr)
 144   (nth 1 (assoc attr attr-list))
 145   )
 146
 147 (defun html2text-get-attr (p1 p2 tag)
 148   (goto-char p1)
 149   (re-search-forward " +[^ ]" p2 t)
 150   (let* ((attr-string (buffer-substring-no-properties (1- (point)) (1- p2)))
 151          (tmp-list (split-string attr-string))
 152          (attr-list)
 153          (counter 0)
 154          (prev (car tmp-list))
 155          (this (nth 1 tmp-list))
 156          (next (nth 2 tmp-list))
 157          (index 1))
 158
 159     (cond
 160      ;; size=3
 161      ((string-match "[^ ]=[^ ]" prev)
 162       (let ((attr  (nth 0 (split-string prev "=")))
 163             (value (nth 1 (split-string prev "="))))
 164         (setq attr-list (cons (list attr value) attr-list))
 165         )
 166       )
 167      ;; size= 3
 168      ((string-match "[^ ]=\\'" prev)
 169       (setq attr-list (cons (list (substring prev 0 -1) this) attr-list))
 170       )
 171      )
 172
 173     (while (< index (length tmp-list))
 174       (cond
 175        ;; size=3
 176        ((string-match "[^ ]=[^ ]" this)
 177         (let ((attr  (nth 0 (split-string this "=")))
 178               (value (nth 1 (split-string this "="))))
 179           (setq attr-list (cons (list attr value) attr-list))
 180           )
 181         )
 182        ;; size =3
 183        ((string-match "\\`=[^ ]" this)
 184         (setq attr-list (cons (list prev (substring this 1)) attr-list)))
 185
 186        ;; size= 3
 187        ((string-match "[^ ]=\\'" this)
 188         (setq attr-list (cons (list (substring this 0 -1) next) attr-list))
 189         )
 190
 191        ;; size = 3
 192        ((string= "=" this)
 193         (setq attr-list (cons (list prev next) attr-list))
 194         )
 195        )
 196       (setq index (1+ index))
 197       (setq prev this)
 198       (setq this next)
 199       (setq next (nth (1+ index) tmp-list))
 200       )
 201
 202     ;;
 203     ;; Tags with no accompanying "=" i.e. value=nil
 204     ;;
 205     (setq prev (car tmp-list))
 206     (setq this (nth 1 tmp-list))
 207     (setq next (nth 2 tmp-list))
 208     (setq index 1)
 209
 210     (if (not (string-match "=" prev))
 211         (progn
 212           (if (not (string= (substring this 0 1) "="))
 213               (setq attr-list (cons (list prev nil) attr-list))
 214             )
 215           )
 216       )
 217
 218     (while (< index (1- (length tmp-list)))
 219       (if (not (string-match "=" this))
 220           (if (not (or (string= (substring next 0 1) "=")
 221                        (string= (substring prev -1) "=")))
 222               (setq attr-list (cons (list this nil) attr-list))
 223             )
 224         )
 225       (setq index (1+ index))
 226       (setq prev this)
 227       (setq this next)
 228       (setq next (nth (1+ index) tmp-list))
 229       )
 230
 231     (if this
 232         (progn
 233           (if (not (string-match "=" this))
 234               (progn
 235                 (if (not (string= (substring prev -1) "="))
 236                     (setq attr-list (cons (list this nil) attr-list))
 237                   )
 238                 )
 239             )
 240           )
 241       )
 242     attr-list ;; return - value
 243     )
 244   )
 245
 246 ;;
 247 ;; </Functions related to attributes>
 248 ;;
 249
 250 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 251 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 252
 253 ;;
 254 ;; <Functions to be called to format a tag-pair>
 255 ;;
 256 (defun html2text-clean-list-items (p1 p2 list-type)
 257   (goto-char p1)
 258   (let ((item-nr 0)
 259         (items   0))
 260     (while (re-search-forward "<li>" p2 t)
 261       (setq items (1+ items)))
 262     (goto-char p1)
 263     (while (< item-nr items)
 264       (setq item-nr (1+ item-nr))
 265       (re-search-forward "<li>" (point-max) t)
 266       (cond
 267        ((string= list-type "ul") (insert " o "))
 268        ((string= list-type "ol") (insert (format " %s: " item-nr)))
 269        (t (insert " x ")))
 270       )
 271     )
 272   )
 273
 274 (defun html2text-clean-dtdd (p1 p2)
 275   (goto-char p1)
 276   (let ((items   0)
 277         (item-nr 0))
 278     (while (re-search-forward "<dt>" p2 t)
 279       (setq items (1+ items)))
 280     (goto-char p1)
 281     (while (< item-nr items)
 282       (setq item-nr (1+ item-nr))
 283       (re-search-forward "<dt>\\([ ]*\\)" (point-max) t)
 284       (when (match-string 1)
 285         (delete-region (point) (- (point) (string-width (match-string 1)))))
 286       (let ((def-p1 (point))
 287             (def-p2 0))
 288         (re-search-forward "\\([ ]*\\)\\(</dt>\\|<dd>\\)" (point-max) t)
 289         (if (match-string 1)
 290             (progn
 291               (let* ((mw1 (string-width (match-string 1)))
 292                      (mw2 (string-width (match-string 2)))
 293                      (mw  (+ mw1 mw2)))
 294                 (goto-char (- (point) mw))
 295                 (delete-region (point) (+ (point) mw1))
 296                 (setq def-p2 (point))))
 297           (setq def-p2 (- (point) (string-width (match-string 2)))))
 298         (put-text-property def-p1 def-p2 'face 'bold)))))
 299
 300 (defun html2text-delete-tags (p1 p2 p3 p4)
 301   (delete-region p1 p2)
 302   (delete-region (- p3 (- p2 p1)) (- p4 (- p2 p1))))
 303
 304 (defun html2text-delete-single-tag (p1 p2)
 305   (delete-region p1 p2))
 306
 307 (defun html2text-clean-hr (p1 p2)
 308   (html2text-delete-single-tag p1 p2)
 309   (goto-char p1)
 310   (newline 1)
 311   (insert (make-string fill-column ?-))
 312   )
 313
 314 (defun html2text-clean-ul (p1 p2 p3 p4)
 315   (html2text-delete-tags p1 p2 p3 p4)
 316   (html2text-clean-list-items p1 (- p3 (- p1 p2)) "ul")
 317   )
 318
 319 (defun html2text-clean-ol (p1 p2 p3 p4)
 320   (html2text-delete-tags p1 p2 p3 p4)
 321   (html2text-clean-list-items p1 (- p3 (- p1 p2)) "ol")
 322   )
 323
 324 (defun html2text-clean-dl (p1 p2 p3 p4)
 325   (html2text-delete-tags p1 p2 p3 p4)
 326   (html2text-clean-dtdd p1 (- p3 (- p1 p2)))
 327   )
 328
 329 (defun html2text-clean-center (p1 p2 p3 p4)
 330   (html2text-delete-tags p1 p2 p3 p4)
 331   (center-region p1 (- p3 (- p2 p1)))
 332   )
 333
 334 (defun html2text-clean-bold (p1 p2 p3 p4)
 335   (put-text-property p2 p3 'face 'bold)
 336   (html2text-delete-tags p1 p2 p3 p4)
 337   )
 338
 339 (defun html2text-clean-title (p1 p2 p3 p4)
 340   (put-text-property p2 p3 'face 'bold)
 341   (html2text-delete-tags p1 p2 p3 p4)
 342   )
 343
 344 (defun html2text-clean-underline (p1 p2 p3 p4)
 345   (put-text-property p2 p3 'face 'underline)
 346   (html2text-delete-tags p1 p2 p3 p4)
 347   )
 348
 349 (defun html2text-clean-italic (p1 p2 p3 p4)
 350   (put-text-property p2 p3 'face 'italic)
 351   (html2text-delete-tags p1 p2 p3 p4)
 352   )
 353
 354 (defun html2text-clean-font (p1 p2 p3 p4)
 355   (html2text-delete-tags p1 p2 p3 p4)
 356   )
 357
 358 (defun html2text-clean-blockquote (p1 p2 p3 p4)
 359   (html2text-delete-tags p1 p2 p3 p4)
 360   )
 361
 362 (defun html2text-clean-anchor (p1 p2 p3 p4)
 363   ;; If someone can explain how to make the URL clickable I will
 364   ;; surely improve upon this.
 365   (let* ((attr-list (html2text-get-attr p1 p2 "a"))
 366          (href (html2text-attr-value attr-list "href")))
 367     (delete-region p1 p4)
 368     (when href
 369       (goto-char p1)
 370       (insert (substring href 1 -1 ))
 371       (put-text-property p1 (point) 'face 'bold))))
 372
 373 ;;
 374 ;; </Functions to be called to format a tag-pair>
 375 ;;
 376
 377 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 378 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 379
 380 ;;
 381 ;; <Functions to be called to fix up paragraphs>
 382 ;;
 383
 384 (defun html2text-fix-paragraph (p1 p2)
 385   (goto-char p1)
 386   (let ((has-br-line)
 387         (refill-start)
 388         (refill-stop))
 389     (if (re-search-forward "<br>$" p2 t)
 390         (setq has-br-line t)
 391       )
 392     (if has-br-line
 393         (progn
 394           (goto-char p1)
 395           (if (re-search-forward ".+[^<][^b][^r][^>]$" p2 t)
 396               (progn
 397                 (beginning-of-line)
 398                 (setq refill-start (point))
 399                 (goto-char p2)
 400                 (re-search-backward ".+[^<][^b][^r][^>]$" refill-start t)
 401                 (next-line 1)
 402                 (end-of-line)
 403                 ;; refill-stop should ideally be adjusted to
 404                 ;; accomodate the "<br>" strings which are removed
 405                 ;; between refill-start and refill-stop.  Can simply
 406                 ;; be returned from my-replace-string
 407                 (setq refill-stop (+ (point)
 408                                      (html2text-replace-string
 409                                       "<br>" ""
 410                                       refill-start (point))))
 411                 ;; (message "Point = %s  refill-stop = %s" (point) refill-stop)
 412                 ;; (sleep-for 4)
 413                 (fill-region refill-start refill-stop)
 414                 )
 415             )
 416           )
 417       )
 418     )
 419   (html2text-replace-string "<br>" "" p1 p2)
 420   )
 421
 422 ;;
 423 ;; This one is interactive ...
 424 ;;
 425 (defun html2text-fix-paragraphs ()
 426   "This _tries_ to fix up the paragraphs - this is done in quite a ad-hook
 427 fashion, quite close to pure guess-work. It does work in some cases though."
 428   (interactive)
 429   (goto-char (point-min))
 430   (replace-regexp "^<br>$" "")
 431   ;; Removing lonely <br> on a single line, if they are left intact we
 432   ;; dont have any paragraphs at all.
 433   (goto-char (point-min))
 434   (while (not (eobp))
 435     (let ((p1 (point)))
 436       (forward-paragraph 1)
 437       ;;(message "Kaller fix med p1=%s  p2=%s " p1 (1- (point))) (sleep-for 5)
 438       (html2text-fix-paragraph p1 (1- (point)))
 439       (goto-char p1)
 440       (when (not (eobp))
 441         (forward-paragraph 1)))))
 442
 443 ;;
 444 ;; </Functions to be called to fix up paragraphs>
 445 ;;
 446
 447 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 448 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
 449
 450 ;;
 451 ;; <Interactive functions>
 452 ;;
 453
 454 (defun html2text-remove-tags (tag-list)
 455   "Removes the tags listed in the list \"html2text-remove-tag-list\".
 456 See the documentation for that variable."
 457   (interactive)
 458   (dolist (tag tag-list)
 459     (goto-char (point-min))
 460     (while (re-search-forward (format "\\(</?%s[^>]*>\\)" tag) (point-max) t)
 461       (delete-region (match-beginning 0) (match-end 0)))))
 462
 463 (defun html2text-format-tags ()
 464   "See the variable \"html2text-format-tag-list\" for documentation"
 465   (interactive)
 466   (dolist (tag-and-function html2text-format-tag-list)
 467     (let ((tag      (car tag-and-function))
 468           (function (cdr tag-and-function)))
 469       (goto-char (point-min))
 470       (while (re-search-forward (format "\\(<%s\\( [^>]*\\)?>\\)" tag)
 471                                 (point-max) t)
 472         (let ((p1)
 473               (p2 (point))
 474               (p3) (p4)
 475               (attr (match-string 1)))
 476           (search-backward "<" (point-min) t)
 477           (setq p1 (point))
 478           (re-search-forward (format "</%s>" tag) (point-max) t)
 479           (setq p4 (point))
 480           (search-backward "</" (point-min) t)
 481           (setq p3 (point))
 482           (funcall function p1 p2 p3 p4)
 483           (goto-char p1)
 484           )
 485         )
 486       )
 487     )
 488   )
 489
 490 (defun html2text-substitute ()
 491   "See the variable \"html2text-replace-list\" for documentation"
 492   (interactive)
 493   (dolist (e html2text-replace-list)
 494     (goto-char (point-min))
 495     (let ((old-string (car e))
 496           (new-string (cdr e)))
 497       (html2text-replace-string old-string new-string (point-min) (point-max))
 498       )
 499     )
 500   )
 501
 502 (defun html2text-format-single-elements ()
 503   ""
 504   (interactive)
 505   (dolist (tag-and-function html2text-format-single-element-list)
 506     (let ((tag      (car tag-and-function))
 507           (function (cdr tag-and-function)))
 508       (goto-char (point-min))
 509       (while (re-search-forward (format "\\(<%s\\( [^>]*\\)?>\\)" tag)
 510                                 (point-max) t)
 511         (let ((p1)
 512               (p2 (point)))
 513           (search-backward "<" (point-min) t)
 514           (setq p1 (point))
 515           (funcall function p1 p2)
 516           )
 517         )
 518       )
 519     )
 520   )
 521
 522 ;;
 523 ;; Main function
 524 ;;
 525
 526 ;;;###autoload
 527 (defun html2text ()
 528   "Convert HTML to plain text in the current buffer."
 529   (interactive)
 530   (save-excursion
 531     (let ((case-fold-search t)
 532           (buffer-read-only))
 533       (html2text-remove-tags html2text-remove-tag-list)
 534       (html2text-format-tags)
 535       (html2text-remove-tags html2text-remove-tag-list2)
 536       (html2text-substitute)
 537       (html2text-format-single-elements)
 538       (html2text-fix-paragraphs))))
 539
 540 ;;
 541 ;; </Interactive functions>
 542 ;;
 543
 544 ;;; arch-tag: e9e57b79-35d4-4de1-a647-e7e01fe56d1e
 545 ;;; html2text.el ends here