lisp/muse-docbook.el

   1 ;;; muse-docbook.el --- publish DocBook files
   2
   3 ;; Copyright (C) 2004, 2005  Free Software Foundation, Inc.
   4
   5 ;; This file is not part of GNU Emacs.
   6
   7 ;; This is free software; you can redistribute it and/or modify it under
   8 ;; the terms of the GNU General Public License as published by the Free
   9 ;; Software Foundation; either version 2, or (at your option) any later
  10 ;; version.
  11 ;;
  12 ;; This is distributed in the hope that it will be useful, but WITHOUT
  13 ;; ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  14 ;; FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  15 ;; for more details.
  16 ;;
  17 ;; You should have received a copy of the GNU General Public License
  18 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  19 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  20 ;; Boston, MA 02110-1301, USA.
  21
  22 ;;; Commentary:
  23
  24 ;;; Contributors:
  25
  26 ;; Dale P. Smith (dpsm AT en DOT com) improved the markup
  27 ;; significantly and made many valuable suggestions.
  28
  29 ;;; Code:
  30
  31 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  32 ;;
  33 ;; Muse DocBook XML Publishing
  34 ;;
  35 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
  36
  37 (require 'muse-publish)
  38 (require 'muse-regexps)
  39
  40 (defgroup muse-docbook nil
  41   "Options controlling the behavior of Muse DocBook XML publishing.
  42 See `muse-docbook' for more information."
  43   :group 'muse-publish)
  44
  45 (defcustom muse-docbook-extension ".xml"
  46   "Default file extension for publishing DocBook XML files."
  47   :type 'string
  48   :group 'muse-docbook)
  49
  50 (defcustom muse-docbook-header
  51   "<?xml version=\"1.0\" encoding=\"<lisp>
  52   (muse-docbook-encoding)</lisp>\"?>
  53 <!DOCTYPE article PUBLIC \"-//OASIS//DTD DocBook V4.2//EN\"
  54                   \"http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd\">
  55 <article>
  56   <articleinfo>
  57     <title><lisp>(muse-publishing-directive \"title\")</lisp></title>
  58     <author><lisp>(muse-docbook-get-author
  59                     (muse-publishing-directive \"author\"))</lisp></author>
  60     <pubdate><lisp>(muse-publishing-directive \"date\")</lisp></pubdate>
  61   </articleinfo>
  62   <!-- Page published by Emacs Muse begins here -->\n"
  63   "Header used for publishing DocBook XML files.
  64 This may be text or a filename."
  65   :type 'string
  66   :group 'muse-docbook)
  67
  68 (defcustom muse-docbook-footer "
  69   <!-- Page published by Emacs Muse ends here -->
  70 </article>\n"
  71   "Footer used for publishing DocBook XML files.
  72 This may be text or a filename."
  73   :type 'string
  74   :group 'muse-docbook)
  75
  76 (defcustom muse-docbook-markup-regexps
  77   `(;; Join together the parts of a table
  78     (10000 ,(concat "  </t\\(body\\|head\\|foot\\)>\\s-*"
  79                     "</tgroup>\\s-*</informaltable>"
  80                     "\\([" muse-regexp-blank "]*\n\\)\\{0,2\\}"
  81                     "[" muse-regexp-blank "]*"
  82                     "<informaltable[^>]*>\\s-*<tgroup[^>]*>\\s-*"
  83                     "<t\\1>\n")
  84            0 "")
  85     (10100 ,(concat "  </tgroup>\\s-*</informaltable>"
  86                     "\\([" muse-regexp-blank "]*\n\\)\\{0,2\\}"
  87                     "[" muse-regexp-blank "]*"
  88                     "<informaltable[^>]*>\\s-*<tgroup[^>]*>\n")
  89            0 "")
  90
  91     ;; Join together the parts of a list
  92     (10200 ,(concat "</\\(itemized\\|ordered\\|variable\\)list>"
  93                     "\\([" muse-regexp-blank "]*\n\\)\\{0,2\\}"
  94                     "[" muse-regexp-blank "]*"
  95                     "<\\1list" "[^>]*>\\s-*")
  96            0 "")
  97
  98     ;; Beginning of doc, end of doc, or plain paragraph separator
  99     (10300 ,(concat "\\(\n</\\(blockquote\\|center\\)>\\)?"
 100                     "\\(\\(\n\\(["
 101                     muse-regexp-blank
 102                     "]*\n\\)+\\)\\|\\`\\s-*\\|\\s-*\\'\\)"
 103                     "\\(<\\(blockquote\\|center\\)>\n\\)?")
 104            0 muse-docbook-markup-paragraph))
 105   "List of markup rules for publishing a Muse page to DocBook XML.
 106 For more on the structure of this list, see `muse-publish-markup-regexps'."
 107   :type '(repeat (choice
 108                   (list :tag "Markup rule"
 109                         integer
 110                         (choice regexp symbol)
 111                         integer
 112                         (choice string function symbol))
 113                   function))
 114   :group 'muse-docbook)
 115
 116 (defcustom muse-docbook-markup-functions
 117   '((anchor . muse-docbook-markup-anchor)
 118     (table . muse-docbook-markup-table))
 119   "An alist of style types to custom functions for that kind of text.
 120 For more on the structure of this list, see
 121 `muse-publish-markup-functions'."
 122   :type '(alist :key-type symbol :value-type function)
 123   :group 'muse-docbook)
 124
 125 (defcustom muse-docbook-markup-strings
 126   '((url-link        . "<ulink url=\"%s\">%s</ulink>")
 127     (internal-link   . "<link linkend=\"%s\">%s</link>")
 128     (email-addr      . "<email>%s</email>")
 129     (emdash          . " &mdash; ")
 130     (rule            . "")
 131     (enddots         . "....")
 132     (dots            . "...")
 133     (section         . "<section><title>")
 134     (section-end     . "</title>")
 135     (subsection      . "<section><title>")
 136     (subsection-end  . "</title>")
 137     (subsubsection   . "<section><title>")
 138     (subsubsection-end . "</title>")
 139     (section-other   . "<section><title>")
 140     (section-other-end . "</title>")
 141     (section-close   . "</section>")
 142     (footnote        . "<footnote><para>")
 143     (footnote-end    . "</para></footnote>")
 144     (begin-underline . "")
 145     (end-underline   . "")
 146     (begin-literal   . "<systemitem>")
 147     (end-literal     . "</systemitem>")
 148     (begin-emph      . "<emphasis>")
 149     (end-emph        . "</emphasis>")
 150     (begin-more-emph . "<emphasis role=\"strong\">")
 151     (end-more-emph   . "</emphasis>")
 152     (begin-most-emph . "<emphasis role=\"strong\"><emphasis>")
 153     (end-most-emph   . "</emphasis></emphasis>")
 154     (begin-verse     . "<literallayout>\n")
 155     (verse-space     . "  ")
 156     (end-verse       . "</literallayout>")
 157     (begin-example   . "<programlisting>")
 158     (end-example     . "</programlisting>")
 159     (begin-center    . "<para role=\"centered\">\n")
 160     (end-center      . "\n</para>")
 161     (begin-quote     . "<blockquote>\n")
 162     (end-quote       . "\n</blockquote>")
 163     (begin-uli       . "<itemizedlist mark=\"bullet\">\n<listitem><para>")
 164     (end-uli         . "</para></listitem>\n</itemizedlist>")
 165     (begin-oli       . "<orderedlist>\n<listitem><para>")
 166     (end-oli         . "</para></listitem>\n</orderedlist>")
 167     (begin-ddt       . "<variablelist>\n<varlistentry>\n<term>")
 168     (start-dde       . "</term>\n<listitem><para>")
 169     (end-ddt         . "</para></listitem>\n</varlistentry>\n</variablelist>"))
 170   "Strings used for marking up text.
 171 These cover the most basic kinds of markup, the handling of which
 172 differs little between the various styles."
 173   :type '(alist :key-type symbol :value-type string)
 174   :group 'muse-docbook)
 175
 176 (defcustom muse-docbook-markup-specials
 177   '((?\" . "&quot;")
 178     (?\< . "&lt;")
 179     (?\> . "&gt;")
 180     (?\& . "&amp;"))
 181   "A table of characters which must be represented specially."
 182   :type '(alist :key-type character :value-type string)
 183   :group 'muse-docbook)
 184
 185 (defcustom muse-docbook-encoding-default 'utf-8
 186   "The default Emacs buffer encoding to use in published files.
 187 This will be used if no special characters are found."
 188   :type 'symbol
 189   :group 'muse-docbook)
 190
 191 (defcustom muse-docbook-charset-default "utf-8"
 192   "The default DocBook XML charset to use if no translation is
 193 found in `muse-docbook-encoding-map'."
 194   :type 'string
 195   :group 'muse-docbook)
 196
 197 (defcustom muse-docbook-encoding-map
 198   '((iso-8859-1         . "iso-8859-1")
 199     (iso-2022-jp        . "iso-2022-jp")
 200     (utf-8              . "utf-8")
 201     (japanese-iso-8bit  . "euc-jp")
 202     (chinese-big5       . "big5")
 203     (mule-utf-8         . "utf-8")
 204     (chinese-iso-8bit   . "gb2312")
 205     (chinese-gbk        . "gbk"))
 206   "An alist mapping emacs coding systems to appropriate DocBook charsets.
 207 Use the base name of the coding system (i.e. without the -unix)."
 208   :type '(alist :key-type coding-system :value-type string)
 209   :group 'muse-docbook)
 210
 211 (defun muse-docbook-transform-content-type (content-type)
 212   "Using `muse-docbook-encoding-map', try and resolve an emacs
 213 coding system to an associated DocBook XML coding system. If no
 214 match is found, `muse-docbook-charset-default' is used instead."
 215   (let ((match (and (fboundp 'coding-system-base)
 216                     (assoc (coding-system-base content-type)
 217                            muse-docbook-encoding-map))))
 218     (if match
 219         (cdr match)
 220       muse-docbook-charset-default)))
 221
 222 (defun muse-docbook-encoding ()
 223   (muse-docbook-transform-content-type
 224    (or (and (boundp 'buffer-file-coding-system)
 225             buffer-file-coding-system)
 226        muse-docbook-encoding-default)))
 227
 228 (defun muse-docbook-markup-paragraph ()
 229   (let ((end (copy-marker (match-end 0) t)))
 230     (goto-char (match-beginning 0))
 231     (when (save-excursion
 232             (save-match-data
 233               (and (re-search-backward "<\\(/?\\)\\(para\\|footnote\\)[ >]"
 234                                        nil t)
 235                    (or (and (string= (match-string 2) "para")
 236                             (not (string= (match-string 1) "/")))
 237                        (and (string= (match-string 2) "footnote")
 238                             (string= (match-string 1) "/"))))))
 239       (insert "</para>"))
 240     (goto-char end))
 241   (cond
 242    ((eobp)
 243     (unless (bolp)
 244       (insert "\n")))
 245    ((eq (char-after) ?\<)
 246     (when (looking-at (concat "<\\(emphasis\\|systemitem"
 247                               "\\|u?link\\|anchor\\|email\\)[ >]"))
 248       (insert "<para>")))
 249    (t
 250     (insert "<para>"))))
 251
 252 (defun muse-docbook-insert-anchor (anchor)
 253   "Insert an anchor, either before the next word, or within a tag."
 254   (skip-chars-forward muse-regexp-space)
 255   (when (looking-at "<\\([^ />]+\\)>")
 256     (goto-char (match-end 0)))
 257   (insert "<anchor id=\"" anchor "\" />"))
 258
 259 (defun muse-docbook-markup-anchor ()
 260   (save-match-data
 261     (muse-docbook-insert-anchor (match-string 1)))
 262   "")
 263
 264 (defun muse-docbook-markup-table ()
 265   (let* ((str (prog1
 266                   (match-string 1)
 267                 (delete-region (match-beginning 0) (match-end 0))))
 268          (fields (split-string str "\\s-*|+\\s-*"))
 269          (type (and (string-match "\\s-*\\(|+\\)\\s-*" str)
 270                     (length (match-string 1 str))))
 271          (part (cond ((= type 1) "tbody")
 272                      ((= type 2) "thead")
 273                      ((= type 3) "tfoot"))))
 274     (insert "<informaltable>\n"
 275             "  <tgroup cols='" (number-to-string (length fields)) "'>\n"
 276             "  <" part ">\n"
 277             "    <row>\n")
 278     (dolist (field fields)
 279       (insert "      <entry>" field "</entry>\n"))
 280     (insert "    </row>\n"
 281             "  </" part ">\n"
 282             "  </tgroup>\n"
 283             "</informaltable>\n")))
 284
 285 (defun muse-docbook-get-author (&optional author)
 286   "Split the AUTHOR directive into separate fields.
 287 AUTHOR should be of the form: \"Firstname Other Names Lastname\",
 288 and anything after `Firstname' is optional."
 289   (setq author (save-match-data (split-string author)))
 290   (let ((num-el (length author)))
 291     (cond ((eq num-el 1)
 292            (concat "<firstname>" (car author) "</firstname>"))
 293           ((eq num-el 2)
 294            (concat "<firstname>" (nth 0 author) "</firstname>"
 295                    "<surname>" (nth 1 author) "</surname>"))
 296           ((eq num-el 3)
 297            (concat "<firstname>" (nth 0 author) "</firstname>"
 298                    "<othername>" (nth 1 author) "</othername>"
 299                    "<surname>" (nth 2 author) "</surname>"))
 300           (t
 301            (let (first last)
 302              (setq first (car author))
 303              (setq author (nreverse (cdr author)))
 304              (setq last (car author))
 305              (setq author (nreverse (cdr author)))
 306              (concat "<firstname>" first "</firstname>"
 307                      "<othername>"
 308                      (mapconcat 'identity author " ")
 309                      "</othername>"
 310                      "<surname>" last "</surname>"))))))
 311
 312 (defun muse-docbook-fixup-tables ()
 313   "Sort table parts."
 314   (goto-char (point-min))
 315   (let (last)
 316     (while (re-search-forward "^ *<tgroup[^>]*>$" nil t)
 317       (unless (get-text-property (point) 'read-only)
 318         (forward-line 1)
 319         (save-restriction
 320           (let ((beg (point)))
 321             (narrow-to-region beg (and (re-search-forward "^ *</tgroup>"
 322                                                           nil t)
 323                                        (match-beginning 0))))
 324           (goto-char (point-min))
 325           (let ((inhibit-read-only t))
 326             (sort-subr nil
 327                        (function
 328                         (lambda ()
 329                           (if (re-search-forward
 330                                "^\\s-*<t\\(head\\|body\\|foot\\)>$" nil t)
 331                               (goto-char (match-beginning 0))
 332                             (goto-char (point-max)))))
 333                        (function
 334                         (lambda ()
 335                           (if (re-search-forward
 336                                "^\\s-*</t\\(head\\|body\\|foot\\)>$" nil t)
 337                               (goto-char (match-end 0))
 338                             (goto-char (point-max)))))
 339                        (function
 340                         (lambda ()
 341                           (looking-at "\\s-*<t\\(head\\|body\\|foot\\)>")
 342                           (cond ((string= (match-string 1) "head") 1)
 343                                 ((string= (match-string 1) "foot") 2)
 344                                 (t 3)))))))))))
 345
 346 (defun muse-docbook-finalize-buffer ()
 347   (when (boundp 'buffer-file-coding-system)
 348     (when (memq buffer-file-coding-system '(no-conversion undecided-unix))
 349       ;; make it agree with the default charset
 350       (setq buffer-file-coding-system muse-docbook-encoding-default))))
 351
 352 ;; Register the Muse DocBook XML Publisher
 353
 354 (unless (assoc "docbook" muse-publishing-styles)
 355   (muse-define-style "docbook"
 356                      :suffix     'muse-docbook-extension
 357                      :regexps    'muse-docbook-markup-regexps
 358                      :functions  'muse-docbook-markup-functions
 359                      :strings    'muse-docbook-markup-strings
 360                      :specials   'muse-docbook-markup-specials
 361                      :before-end 'muse-docbook-fixup-tables
 362                      :after      'muse-docbook-finalize-buffer
 363                      :header     'muse-docbook-header
 364                      :footer     'muse-docbook-footer
 365                      :browser    'find-file))
 366
 367 (provide 'muse-docbook)
 368
 369 ;;; muse-docbook.el ends here