lisp/dom.el

   1 ;;; dom.el --- XML/HTML (etc.) DOM manipulation and searching functions
   2
   3 ;; Copyright (C) 2014-2015 Free Software Foundation, Inc.
   4
   5 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
   6 ;; Keywords: xml, html
   7
   8 ;; This file is part of GNU Emacs.
   9
  10 ;; GNU Emacs is free software: you can redistribute it and/or modify
  11 ;; it under the terms of the GNU General Public License as published by
  12 ;; the Free Software Foundation, either version 3 of the License, or
  13 ;; (at your option) any later version.
  14
  15 ;; GNU Emacs is distributed in the hope that it will be useful,
  16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 ;; GNU General Public License for more details.
  19
  20 ;; You should have received a copy of the GNU General Public License
  21 ;; along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.
  22
  23 ;;; Commentary:
  24
  25 ;;; Code:
  26
  27 (require 'cl-lib)
  28
  29 (defsubst dom-tag (node)
  30   "Return the NODE tag."
  31   ;; Called on a list of nodes.  Use the first.
  32   (if (consp (car node))
  33       (caar node)
  34     (car node)))
  35
  36 (defsubst dom-attributes (node)
  37   "Return the NODE attributes."
  38   ;; Called on a list of nodes.  Use the first.
  39   (if (consp (car node))
  40       (cadr (car node))
  41     (cadr node)))
  42
  43 (defsubst dom-children (node)
  44   "Return the NODE children."
  45   ;; Called on a list of nodes.  Use the first.
  46   (if (consp (car node))
  47       (cddr (car node))
  48     (cddr node)))
  49
  50 (defun dom-non-text-children (node)
  51   "Return all non-text-node children of NODE."
  52   (cl-loop for child in (dom-children node)
  53            unless (stringp child)
  54            collect child))
  55
  56 (defun dom-set-attributes (node attributes)
  57   "Set the attributes of NODE to ATTRIBUTES."
  58   (setq node (dom-ensure-node node))
  59   (setcar (cdr node) attributes))
  60
  61 (defun dom-set-attribute (node attribute value)
  62   "Set ATTRIBUTE in NODE to VALUE."
  63   (setq node (dom-ensure-node node))
  64   (let ((old (assoc attribute (cadr node))))
  65     (if old
  66         (setcdr old value)
  67       (setcar (cdr node) (nconc (cadr node) (list (cons attribute value)))))))
  68
  69 (defmacro dom-attr (node attr)
  70   "Return the attribute ATTR from NODE.
  71 A typical attribute is `href'."
  72   `(cdr (assq ,attr (dom-attributes ,node))))
  73
  74 (defun dom-text (node)
  75   "Return all the text bits in the current node concatenated."
  76   (mapconcat 'identity (cl-remove-if-not 'stringp (dom-children node)) " "))
  77
  78 (defun dom-texts (node &optional separator)
  79   "Return all textual data under NODE concatenated with SEPARATOR in-between."
  80   (mapconcat
  81    'identity
  82    (mapcar
  83     (lambda (elem)
  84       (if (stringp elem)
  85           elem
  86         (dom-texts elem separator)))
  87     (dom-children node))
  88    (or separator " ")))
  89
  90 (defun dom-child-by-tag (dom tag)
  91   "Return the first child of DOM that is of type TAG."
  92   (assoc tag (dom-children dom)))
  93
  94 (defun dom-by-tag (dom tag)
  95   "Return elements in DOM that is of type TAG.
  96 A name is a symbol like `td'."
  97   (let ((matches (cl-loop for child in (dom-children dom)
  98                           for matches = (and (not (stringp child))
  99                                              (dom-by-tag child tag))
 100                           when matches
 101                           append matches)))
 102     (if (equal (dom-tag dom) tag)
 103         (cons dom matches)
 104       matches)))
 105
 106 (defun dom-strings (dom)
 107   "Return elements in DOM that are strings."
 108   (cl-loop for child in (dom-children dom)
 109            if (stringp child)
 110            collect child
 111            else
 112            append (dom-strings child)))
 113
 114 (defun dom-by-class (dom match)
 115   "Return elements in DOM that have a class name that matches regexp MATCH."
 116   (dom-elements dom 'class match))
 117
 118 (defun dom-by-style (dom match)
 119   "Return elements in DOM that have a style that matches regexp MATCH."
 120   (dom-elements dom 'style match))
 121
 122 (defun dom-by-id (dom match)
 123   "Return elements in DOM that have an ID that matches regexp MATCH."
 124   (dom-elements dom 'id match))
 125
 126 (defun dom-elements (dom attribute match)
 127   "Find elements matching MATCH (a regexp) in ATTRIBUTE.
 128 ATTRIBUTE would typically be `class', `id' or the like."
 129   (let ((matches (cl-loop for child in (dom-children dom)
 130                           for matches = (and (not (stringp child))
 131                                              (dom-elements child attribute
 132                                                            match))
 133                           when matches
 134                           append matches))
 135         (attr (dom-attr dom attribute)))
 136     (if (and attr
 137              (string-match match attr))
 138         (cons dom matches)
 139       matches)))
 140
 141 (defun dom-parent (dom node)
 142   "Return the parent of NODE in DOM."
 143   (if (memq node (dom-children dom))
 144       dom
 145     (let ((result nil))
 146       (dolist (elem (dom-children dom))
 147         (when (and (not result)
 148                    (not (stringp elem)))
 149           (setq result (dom-parent elem node))))
 150       result)))
 151
 152 (defun dom-previous-sibling (dom node)
 153   (when-let (parent (dom-parent dom node))
 154     (let ((siblings (dom-children parent))
 155           (previous nil))
 156       (while siblings
 157         (when (eq (cadr siblings) node)
 158           (setq previous (car siblings)))
 159         (pop siblings))
 160       previous)))
 161
 162 (defun dom-node (tag &optional attributes &rest children)
 163   "Return a DOM node with TAG and ATTRIBUTES."
 164   (if children
 165       `(,tag ,attributes ,@children)
 166     (list tag attributes)))
 167
 168 (defun dom-append-child (node child)
 169   "Append CHILD to the end of NODE's children."
 170   (setq node (dom-ensure-node node))
 171   (nconc node (list child)))
 172
 173 (defun dom-add-child-before (node child &optional before)
 174   "Add CHILD to NODE's children before child BEFORE.
 175 If BEFORE is nil, make CHILD NODE's first child."
 176   (setq node (dom-ensure-node node))
 177   (let ((children (dom-children node)))
 178     (when (and before
 179                (not (memq before children)))
 180       (error "%s does not exist as a child" before))
 181     (let ((pos (if before
 182                    (cl-position before children)
 183                  0)))
 184       (if (zerop pos)
 185           ;; First child.
 186           (setcdr (cdr node) (cons child (cddr node)))
 187         (setcdr (nthcdr (1- pos) children)
 188                 (cons child (nthcdr pos children))))))
 189   node)
 190
 191 (defun dom-ensure-node (node)
 192   "Ensure that NODE is a proper DOM node."
 193   ;; Add empty attributes, if none.
 194   (when (consp (car node))
 195     (setq node (car node)))
 196   (when (= (length node) 1)
 197     (setcdr node (list nil)))
 198   node)
 199
 200 (defun dom-pp (dom &optional remove-empty)
 201   "Pretty-print DOM at point.
 202 If REMOVE-EMPTY, ignore textual nodes that contain just
 203 white-space."
 204   (let ((column (current-column)))
 205     (insert (format "(%S " (dom-tag dom)))
 206     (let* ((attr (dom-attributes dom))
 207            (times (length attr))
 208            (column (1+ (current-column))))
 209       (if (null attr)
 210           (insert "nil")
 211         (insert "(")
 212         (dolist (elem attr)
 213           (insert (format "(%S . %S)" (car elem) (cdr elem)))
 214           (if (zerop (cl-decf times))
 215               (insert ")")
 216             (insert "\n" (make-string column ? ))))))
 217     (let* ((children (if remove-empty
 218                          (cl-remove-if
 219                           (lambda (child)
 220                             (and (stringp child)
 221                                  (string-match "\\`[\n\r\t  ]*\\'" child)))
 222                           (dom-children dom))
 223                        (dom-children dom)))
 224            (times (length children)))
 225       (if (null children)
 226           (insert ")")
 227         (insert "\n" (make-string (1+ column) ? ))
 228         (dolist (child children)
 229           (if (stringp child)
 230               (if (or (not remove-empty)
 231                       (not (string-match "\\`[\n\r\t  ]*\\'" child)))
 232                   (insert (format "%S" child)))
 233             (dom-pp child remove-empty))
 234           (if (zerop (cl-decf times))
 235               (insert ")")
 236             (insert "\n" (make-string (1+ column) ? ))))))))
 237
 238 (provide 'dom)
 239
 240 ;;; dom.el ends here