4 Copyright 1999,2000,2001,2002,2004,2005 $ThePhpWikiProgrammingTeam
6 This file is part of PhpWiki.
8 PhpWiki 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 of the License, or
11 (at your option) any later version.
13 PhpWiki 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.
18 You should have received a copy of the GNU General Public License
19 along with PhpWiki; if not, write to the Free Software
20 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
23 require_once('lib/TextSearchQuery.php');
24 require_once("lib/PageList.php");
27 * Case insensitive fulltext search
28 * Options: case_exact, regex, hilight
31 * TODO: Hooks to search in external documents: ExternalTextSearch
32 * Only uploaded: textfiles, PDF, HTML, DOC, XLS, ... or
33 * External apps: xapian-omages seems to be the best, over lucene.net,
34 * swish, nakamazu, ...
36 * See http://sf.net/tracker/index.php?aid=927395&group_id=6121&atid=106121
37 * Wordaround to let the dead locks occur somewhat later:
38 * increased the memory limit of PHP4 from 8 MB to 32 MB
39 * php.ini: memory_limit = 32 MB
41 class WikiPlugin_FullTextSearch
45 return _("FullTextSearch");
48 function getDescription() {
49 return _("Search the content of all pages in this wiki.");
52 function getVersion() {
53 return preg_replace("/[Revision: $]/", '',
57 function getDefaultArguments() {
60 PageList::supportedArgs(), // paging and more.
63 'case_exact' => false,
65 'sortby' => '-hi_content',
67 'exclude' => false, // comma-seperated list of glob
68 'quiet' => true)); // be less verbose
71 function run($dbi, $argstr, &$request, $basepage) {
73 $args = $this->getArgs($argstr, $request);
74 if (empty($args['s']))
78 $query = new TextSearchQuery($s, $case_exact, $regex);
79 $pages = $dbi->fullSearch($query, $sortby, $limit, $exclude);
81 $hilight_re = $hilight ? $query->getHighlightRegexp() : false;
84 if ($quiet) { // see how easy it is with PageList...
86 $args['listtype'] = 'dl';
87 $args['types'] = array(new _PageList_Column_content
88 ('rev:hi_content', _("Content"), "left", $s));
89 $list = new PageList(false, $exclude, $args);
90 $list->setCaption(fmt("Full text search results for '%s'", $s));
91 while ($page = $pages->next()) {
92 $list->addPage( $page );
97 // Todo: we should better define a new PageListDL class for dl/dt/dd lists
98 // But the new column types must have a callback then. (showhits)
99 // See e.g. WikiAdminSearchReplace for custom pagelist columns
101 if (!$limit or !is_int($limit))
103 // expand all page wildcards to a list of pages which should be ignored
104 if ($exclude) $exclude = explodePageList($exclude);
105 while ($page = $pages->next() and (!$limit or ($count < $limit))) {
106 $name = $page->getName();
107 if ($exclude and in_array($name,$exclude)) continue;
109 $list->pushContent(HTML::dt(WikiLink($page)));
111 $list->pushContent($this->showhits($page, $hilight_re));
114 if ($limit and $count >= $limit) //todo: pager link to list of next matches
115 $list->pushContent(HTML::dd(fmt("only %d pages displayed",$limit)));
116 if (!$list->getContent())
117 $list->pushContent(HTML::dd(_("<no matches>")));
119 if (!empty($pages->stoplisted))
120 $list = HTML(HTML::p(fmt(_("Ignored stoplist words '%s'"),
121 join(', ', $pages->stoplisted))),
125 return HTML(HTML::p(fmt("Full text search results for '%s'", $s)),
129 function showhits($page, $hilight_re) {
130 $current = $page->getCurrentRevision();
131 $matches = preg_grep("/$hilight_re/i", $current->getContent());
133 foreach ($matches as $line) {
134 $line = $this->highlight_line($line, $hilight_re);
135 $html[] = HTML::dd(HTML::small(array('class' => 'search-context'),
141 function highlight_line ($line, $hilight_re) {
142 while (preg_match("/^(.*?)($hilight_re)/i", $line, $m)) {
143 $line = substr($line, strlen($m[0]));
144 $html[] = $m[1]; // prematch
145 $html[] = HTML::strong(array('class' => 'search-term'), $m[2]); // match
147 $html[] = $line; // postmatch
153 * List of Links and link to ListLinks
155 class _PageList_Column_hilight extends _PageList_Column {
156 function _PageList_Column_WantedPages_links (&$params) {
157 $this->parentobj =& $params[3];
158 $this->_PageList_Column($params[0],$params[1],$params[2]);
160 function _getValue(&$page, $revision_handle) {
162 $pagename = $page->getName();
163 $count = count($this->parentobj->_wpagelist[$pagename]);
164 return LinkURL(WikiURL($page, array('action' => 'BackLinks'), false),
165 fmt("(%d Links)", $count));
170 // $Log: not supported by cvs2svn $
171 // Revision 1.28 2007/06/07 17:02:42 rurban
172 // fix display of pagenames containing ":" in certain lists
174 // Revision 1.27 2007/01/04 16:46:40 rurban
177 // Revision 1.26 2005/11/14 22:33:04 rurban
178 // print ignored stoplist words
180 // Revision 1.25 2005/09/11 14:55:05 rurban
181 // implement fulltext stoplist
183 // Revision 1.24 2004/11/26 18:39:02 rurban
184 // new regex search parser and SQL backends (90% complete, glob and pcre backends missing)
186 // Revision 1.23 2004/11/23 15:17:19 rurban
187 // better support for case_exact search (not caseexact for consistency),
188 // plugin args simplification:
189 // handle and explode exclude and pages argument in WikiPlugin::getArgs
190 // and exclude in advance (at the sql level if possible)
191 // handle sortby and limit from request override in WikiPlugin::getArgs
192 // ListSubpages: renamed pages to maxpages
194 // Revision 1.22 2004/05/28 11:01:58 rurban
195 // support to disable highlighting
196 // example: s=ReiniUrban&hilight=0&noheader=1
198 // Revision 1.21 2004/04/18 01:11:52 rurban
199 // more numeric pagename fixes.
200 // fixed action=upload with merge conflict warnings.
201 // charset changed from constant to global (dynamic utf-8 switching)
203 // Revision 1.20 2004/02/28 21:14:08 rurban
204 // generally more PHPDOC docs
205 // see http://xarch.tu-graz.ac.at/home/rurban/phpwiki/xref/
206 // fxied WikiUserNew pref handling: empty theme not stored, save only
207 // changed prefs, sql prefs improved, fixed password update,
208 // removed REPLACE sql (dangerous)
209 // moved gettext init after the locale was guessed
210 // + some minor changes
212 // Revision 1.19 2004/02/26 04:27:39 rurban
213 // wrong limit notification
215 // Revision 1.18 2004/02/26 04:24:03 rurban
216 // simplify quiet handling by using PageList
218 // Revision 1.17 2004/02/26 04:03:39 rurban
219 // added quiet, limit and exclude to FullTextSearch,
220 // fixed explodePageList with previously unloaded PageList
222 // Revision 1.16 2004/02/17 12:11:36 rurban
223 // added missing 4th basepage arg at plugin->run() to almost all plugins. This caused no harm so far, because it was silently dropped on normal usage. However on plugin internal ->run invocations it failed. (InterWikiSearch, IncludeSiteMap, ...)
225 // Revision 1.15 2003/01/18 21:41:01 carstenklapp
227 // Reformatting & tabs to spaces;
228 // Added copyleft, getVersion, getDescription, rcs_id.
235 // c-hanging-comment-ender-p: nil
236 // indent-tabs-mode: nil