3 * StatusNet - the distributed open-source microblogging tool
4 * Copyright (C) 2010, StatusNet, Inc.
6 * Importer class for Delicious.com backups
10 * This program is free software: you can redistribute it and/or modify
11 * it under the terms of the GNU Affero 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.
15 * This program 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 Affero General Public License for more details.
20 * You should have received a copy of the GNU Affero General Public License
21 * along with this program. If not, see <http://www.gnu.org/licenses/>.
25 * @author Evan Prodromou <evan@status.net>
26 * @copyright 2010 StatusNet, Inc.
27 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html AGPL 3.0
28 * @link http://status.net/
31 if (!defined('STATUSNET')) {
32 // This check helps protect against security problems;
33 // your code file can't be executed directly from the web.
38 * Importer class for Delicious bookmarks
42 * @author Evan Prodromou <evan@status.net>
43 * @copyright 2010 StatusNet, Inc.
44 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html AGPL 3.0
45 * @link http://status.net/
48 class DeliciousBackupImporter extends QueueHandler
51 * Transport of the importer
53 * @return string transport string
62 * Import an in-memory bookmark list to a user's account
64 * Take a delicious.com backup file (same as Netscape bookmarks.html)
65 * and import to StatusNet as Bookmark activities.
67 * The document format is terrible. It consists of a <dl> with
68 * a bunch of <dt>'s, occasionally with <dd>'s adding descriptions.
69 * There are sometimes <p>'s lost inside.
71 * @param array $data pair of user, text
73 * @return boolean success value
76 function handle($data)
78 list($user, $body) = $data;
80 $doc = $this->importHTML($body);
82 // If we can't parse it, it's no good
88 $dls = $doc->getElementsByTagName('dl');
90 if ($dls->length != 1) {
91 throw new ClientException(_("Bad import file."));
96 $children = $dl->childNodes;
100 for ($i = 0; $i < $children->length; $i++) {
102 $child = $children->item($i);
103 if ($child->nodeType != XML_ELEMENT_NODE) {
106 switch (strtolower($child->tagName)) {
108 // <dt> nodes contain primary information about a bookmark.
109 // We can't import the current one just yet though, since
110 // it may be followed by a <dd>.
113 $this->importBookmark($user, $dt);
122 // This <dd> contains a description for the bookmark in
123 // the preceding <dt> node.
124 $saved = $this->importBookmark($user, $dt, $dd);
131 common_log(LOG_INFO, 'Skipping the <p> in the <dl>.');
134 common_log(LOG_WARNING,
135 "Unexpected element $child->tagName ".
136 " found in import.");
138 } catch (Exception $e) {
139 common_log(LOG_ERR, $e->getMessage());
144 // There was a final bookmark without a description.
146 $this->importBookmark($user, $dt);
147 } catch (Exception $e) {
148 common_log(LOG_ERR, $e->getMessage());
156 * Import a single bookmark
158 * Takes a <dt>/<dd> pair. The <dt> has a single
159 * <a> in it with some non-standard attributes.
161 * A <dt><dt><dd> sequence will appear as a <dt> with
162 * anothe <dt> as a child. We handle this case recursively.
164 * @param User $user User to import data as
165 * @param DOMElement $dt <dt> element
166 * @param DOMElement $dd <dd> element
168 * @return Notice imported notice
171 function importBookmark($user, $dt, $dd = null)
173 $as = $dt->getElementsByTagName('a');
175 if ($as->length == 0) {
176 throw new ClientException(_("No <A> tag in a <DT>."));
181 $private = $a->getAttribute('private');
184 throw new ClientException(_('Skipping private bookmark.'));
188 $description = $dd->nodeValue;
192 $addDate = $a->getAttribute('add_date');
195 'profile_id' => $user->id,
196 'title' => $a->nodeValue,
197 'description' => $description,
198 'url' => $a->getAttribute('href'),
199 'tags' => $a->getAttribute('tags'),
200 'created' => common_sql_date(intval($addDate))
203 $qm = QueueManager::get();
204 $qm->enqueue($data, 'dlcsbkmk');
210 * Hides the errors that the dom parser returns
212 * @param string $body Data to import
214 * @return DOMDocument parsed document
217 function importHTML($body)
219 // DOMDocument::loadHTML may throw warnings on unrecognized elements,
220 // and notices on unrecognized namespaces.
221 $old = error_reporting(error_reporting() & ~(E_WARNING | E_NOTICE));
222 $dom = new DOMDocument();
223 $ok = $dom->loadHTML($body);
224 error_reporting($old);
227 foreach ($dom->getElementsByTagName('body') as $node) {
228 $this->fixListsIn($node);
237 function fixListsIn(DOMNode $body) {
240 foreach ($body->childNodes as $node) {
241 if ($node->nodeType == XML_ELEMENT_NODE) {
242 $el = strtolower($node->nodeName);
249 foreach ($toFix as $node) {
250 $this->fixList($node);
254 function fixList(DOMNode $list) {
257 foreach ($list->childNodes as $node) {
258 if ($node->nodeType == XML_ELEMENT_NODE) {
259 $el = strtolower($node->nodeName);
260 if ($el == 'dt' || $el == 'dd') {
265 // Technically, these can only appear inside a <dd>...
266 $this->fixList($node);
271 foreach ($toFix as $node) {
272 $this->fixListItem($node);
276 function fixListItem(DOMNode $item) {
277 // The HTML parser in libxml2 doesn't seem to properly handle
278 // many cases of implied close tags, apparently because it doesn't
279 // understand the nesting rules specified in the HTML DTD.
281 // This leads to sequences of adjacent <dt>s or <dd>s being incorrectly
282 // interpreted as parent->child trees instead of siblings:
284 // When parsing this input: "<dt>aaa <dt>bbb"
285 // should be equivalent to: "<dt>aaa </dt><dt>bbb</dt>"
286 // but we're seeing instead: "<dt>aaa <dt>bbb</dt></dt>"
288 // It does at least know that going from dt to dd, or dd to dt,
289 // should make a break.
293 foreach ($item->childNodes as $node) {
294 if ($node->nodeType == XML_ELEMENT_NODE) {
295 $el = strtolower($node->nodeName);
296 if ($el == 'dt' || $el == 'dd') {
297 // dt & dd cannot contain each other;
298 // This node was incorrectly placed; move it up a level!
303 // Technically, these can only appear inside a <dd>.
304 $this->fixList($node);
309 $parent = $item->parentNode;
310 $next = $item->nextSibling;
311 foreach ($toMove as $node) {
312 $item->removeChild($node);
313 $parent->insertBefore($node, $next);
314 $this->fixListItem($node);